{ "generated_at": "2026-09-04T10:48:22+02:00", "research_model": "gpt-5.6-terra", "providers": [ { "id": "anthropic", "name": "Anthropic", "notes": "Die Basispreise gelten pro 1 Mio. Tokens; Prompt Caching und Batch-Verarbeitung haben abweichende Tarife. Claude Pro und Max sind Abonnements für Claude-Apps, keine separaten API-Modellvarianten. Bei Legacy-Modellen sind gemäß Vorgabe nur die jeweils unmittelbar abgelösten Vorgänger erfasst.", "pricing_url": "https://platform.claude.com/docs/en/about-claude/pricing", "docs_url": "https://platform.claude.com/docs/en/models/overview", "researched_at": "2026-09-04T10:44:38+0200", "usage": { "input_tokens": 109443, "output_tokens": 8549, "web_searches": 11 }, "model_count": 9, "live_checked": true, "live_error": null }, { "id": "xai", "name": "xAI (Grok)", "notes": "Stand 2026-09-04 führt die offizielle xAI-Preisseite sieben gehostete Text-/Bild-zu-Text-Sprachmodelle auf. Die Standardpreise gelten unter 200.000 Prompt-Tokens; ab diesem Schwellenwert gilt für sämtliche Tokens eines Requests der jeweilige Long-Context-Tarif. Priority Processing kostet laut Preisseite das Doppelte des Standardtarifs; Batch ist nur bei den entsprechend gekennzeichneten Modellen verfügbar und dort um 20 % reduziert.", "pricing_url": "https://docs.x.ai/developers/pricing", "docs_url": "https://docs.x.ai/developers/models", "researched_at": "2026-09-04T10:44:45+0200", "usage": { "input_tokens": 101147, "output_tokens": 8575, "web_searches": 11 }, "model_count": 8, "live_checked": false, "live_error": null }, { "id": "openai", "name": "OpenAI", "notes": "Stand: 2026-09-04. Die reguläre API-Produktlinie besteht aus GPT-6 Astra (limitiert ausgerollt), der GPT-5.6-Familie sowie weiterhin buchbaren GPT-5.5-/GPT-5.4-Vorgängern und GPT-5.3-Codex. Batch und Flex sind bei den aktuellen GPT-5.6-Modellen günstiger, Fast Mode teurer; lange Kontexte können einen Aufschlag auslösen.", "pricing_url": "https://platform.openai.com/docs/pricing", "docs_url": "https://platform.openai.com/docs/models", "researched_at": "2026-09-04T10:44:48+0200", "usage": { "input_tokens": 100820, "output_tokens": 8924, "web_searches": 10 }, "model_count": 14, "live_checked": true, "live_error": null }, { "id": "mistral", "name": "Mistral AI", "notes": "Stand: 2026-09-04. Standardpreise gelten für die globale Serverless-API; Batch-Verarbeitung halbiert laut Anbieter die Tokenpreise, regionale Inferenz kann für unterstützte Modelle 10 % Aufschlag haben. Die fünf als „legacy“ geführten Vorgänger sind nach der dokumentierten sechsmonatigen Deprecation-Frist am Stichtag noch erreichbar; Labs-Vorgänger mit nur einem Monat Frist wurden nicht aufgenommen.", "pricing_url": "https://mistral.ai/pricing/api/", "docs_url": "https://docs.mistral.ai/models", "researched_at": "2026-09-04T10:44:56+0200", "usage": { "input_tokens": 71094, "output_tokens": 10155, "web_searches": 7 }, "model_count": 15, "live_checked": false, "live_error": null }, { "id": "google", "name": "Google (Gemini)", "notes": "Google bietet über die Gemini API einen kostenlosen Einstieg sowie kostenpflichtige Standard-, Batch-, Flex- und Priority-Inference an. Batch und Flex sind bei den meisten Tokenmodellen günstiger als Standard; Priority ist teurer. Preise können nach Kontextlänge oder Eingabemodalität variieren und sind je Modell in den Preisnotizen zusammengefasst.", "pricing_url": "https://ai.google.dev/gemini-api/docs/pricing", "docs_url": "https://ai.google.dev/gemini-api/docs/models", "researched_at": "2026-09-04T10:45:13+0200", "usage": { "input_tokens": 116812, "output_tokens": 12240, "web_searches": 12 }, "model_count": 18, "live_checked": true, "live_error": null }, { "id": "meta", "name": "Meta (Llama)", "notes": "Die frühere Llama API wurde durch die Meta Model API abgelöst; diese befindet sich am 4. September 2026 weiterhin im Public Preview. Muse Spark 1.3 ist das aktuelle proprietäre API-Flaggschiff; die günstigere „Contributor“-Route ist derselbe Modellcheckpoint, erlaubt Meta jedoch die Nutzung von Ein- und Ausgaben zur Produktverbesserung. Llama 4 und Muse Glimmer sind Open-Weights-Angebote ohne direkte, von Meta betriebene Token-API.", "pricing_url": "https://dev.meta.ai/docs/models", "docs_url": "https://dev.meta.ai/docs/models", "researched_at": "2026-09-04T10:45:31+0200", "usage": { "input_tokens": 104870, "output_tokens": 9101, "web_searches": 23 }, "model_count": 6, "live_checked": false, "live_error": null }, { "id": "deepseek", "name": "DeepSeek", "notes": "Stand 2026-09-04 führt DeepSeek für seine API drei Chatmodell-IDs; die früheren IDs `deepseek-chat` und `deepseek-reasoner` wurden am 2026-07-24 um 15:59 UTC eingestellt und sind daher keine Legacy-Einträge mehr. Die unten als Standardpreise verwendeten Peak-Tarife gelten werktags 01:00–04:00 sowie 06:00–10:00 UTC; alle übrigen Zeiten werden mit 50 % Rabatt abgerechnet.", "pricing_url": "https://api-docs.deepseek.com/quick_start/pricing/", "docs_url": "https://api-docs.deepseek.com/", "researched_at": "2026-09-04T10:45:45+0200", "usage": { "input_tokens": 58608, "output_tokens": 5269, "web_searches": 7 }, "model_count": 3, "live_checked": true, "live_error": null }, { "id": "moonshot", "name": "Moonshot AI (Kimi)", "notes": "Moonshot AI bietet auf der Kimi API Platform nutzungsbasierte Abrechnung pro Token; die angegebenen Preise verstehen sich ohne anfallende Steuern. Automatisches Context Caching senkt den Preis für Cache-Hits. Der Batch-Modus kostet 60 % des Echtzeit-Tarifs, unterstützt aber nur kimi-k2.7-code und kimi-k2.6, nicht K3.", "pricing_url": "https://platform.kimi.ai/docs/pricing/chat", "docs_url": "https://platform.kimi.ai/docs/models", "researched_at": "2026-09-04T10:46:20+0200", "usage": { "input_tokens": 113193, "output_tokens": 5254, "web_searches": 16 }, "model_count": 4, "live_checked": true, "live_error": null }, { "id": "minimax", "name": "MiniMax", "notes": "MiniMax bietet aktuell ein 1M-kontextiges, multimodales Flaggschiff (M3), die weiterhin regulär bepreisten M2.7-Varianten sowie ältere M2.5-Varianten an. Die API-Standardpreise gelten pro 1 Mio. Tokens; M3 wird oberhalb von 512k Input-Tokens teurer, und der Priority-Service kostet 1,5× des Standardtarifs. M2-her ist weiterhin per API dokumentiert, steht jedoch nicht mehr mit einem separaten Tarif in der aktuellen Pay-as-you-go-Tabelle.", "pricing_url": "https://platform.minimax.io/docs/guides/pricing-paygo", "docs_url": "https://platform.minimax.io/docs/guides/text-generation", "researched_at": "2026-09-04T10:46:56+0200", "usage": { "input_tokens": 118423, "output_tokens": 5313, "web_searches": 14 }, "model_count": 6, "live_checked": false, "live_error": null }, { "id": "cohere", "name": "Cohere", "notes": "Stand 4. September 2026. Cohere rechnet reguläre API-Modelle grundsätzlich tokenbasiert ab; mehrere neuere Modelle sind über die Standard-API zunächst nur rate-limitiert kostenlos nutzbar und erfordern für produktive Nutzung Model Vault bzw. ein Sales-Angebot. Datierte Cohere-Modellversionen wurden gemäß Vorgabe zu undatierten Familien-IDs normalisiert; die konkreten, aktuell dokumentierten Snapshots stehen in den Quellseiten.", "pricing_url": "https://cohere.com/pricing", "docs_url": "https://docs.cohere.com/docs/models", "researched_at": "2026-09-04T10:47:09+0200", "usage": { "input_tokens": 109170, "output_tokens": 10228, "web_searches": 12 }, "model_count": 17, "live_checked": false, "live_error": null }, { "id": "amazon", "name": "Amazon (Nova)", "notes": "Amazon Nova wird über Amazon Bedrock nutzungsbasiert pro Token abgerechnet; die Tarifseite unterscheidet je nach Modell zwischen Standard-, Flex-, Priority- und Batch-Inferenz. Nova 2 ist die aktuelle Generation; die weiterhin API-verfügbaren Nova-1-Modelle sind hier als „legacy“ eingeordnet. Nicht enthalten sind Sonic (Live-Sprache), Embeddings, Canvas/Reel sowie Nova Act.", "pricing_url": "https://aws.amazon.com/bedrock/pricing/", "docs_url": "https://docs.aws.amazon.com/nova/latest/nova2-userguide/what-is-nova-2.html", "researched_at": "2026-09-04T10:47:17+0200", "usage": { "input_tokens": 122893, "output_tokens": 8613, "web_searches": 16 }, "model_count": 7, "live_checked": false, "live_error": null }, { "id": "zai", "name": "Z.ai / Zhipu (GLM)", "notes": "Stand 2026-09-04. Z.AI rechnet API-Nutzung pro 1 Mio. Tokens ab; bei den kostenpflichtigen Chatmodellen ist Context-Cache-Speicherung derzeit zeitlich begrenzt kostenlos. GLM-5.3-Flash läuft aktuell zu einem bis 2026-09-09 (UTC+8) befristet um 50 % reduzierten Tarif; die unten stehenden Preise sind die aktuell berechneten Preise. Die noch angebotenen GLM-4.5-/4.5V-Modelle wurden als weiter zurückliegende Historie nicht aufgenommen; enthalten ist jeweils nur die letzte abgelöste GLM-4.6-/4.6V-Generation.", "pricing_url": "https://docs.z.ai/guides/overview/pricing", "docs_url": "https://docs.z.ai/api-reference/llm/chat-completion", "researched_at": "2026-09-04T10:47:22+0200", "usage": { "input_tokens": 112968, "output_tokens": 9856, "web_searches": 13 }, "model_count": 12, "live_checked": false, "live_error": null }, { "id": "perplexity", "name": "Perplexity (Sonar)", "notes": "Die Sonar API ist in der aktuellen Perplexity-Dokumentation als „Legacy API“ eingeordnet, ihre vier dokumentierten Modelle sind jedoch weiterhin über den Sonar-Chat-Completion-Endpunkt nutzbar und auf der Preisseite ausgewiesen. Abrechnung erfolgt nutzungsbasiert ohne Abonnement; neben Tokenpreisen fallen bei Sonar, Sonar Pro und Sonar Reasoning Pro kontextabhängige Request-Gebühren an.", "pricing_url": "https://docs.perplexity.ai/docs/getting-started/pricing", "docs_url": "https://docs.perplexity.ai/docs/sonar/models", "researched_at": "2026-09-04T10:47:23+0200", "usage": { "input_tokens": 77254, "output_tokens": 4652, "web_searches": 8 }, "model_count": 4, "live_checked": false, "live_error": null }, { "id": "alibaba", "name": "Alibaba (Qwen)", "notes": "Abgerechnet wird bei Alibaba Cloud Model Studio standardmäßig nutzungsbasiert pro Token; die Preise unterscheiden sich je nach Region, Eingabelänge, Thinking-Modus und Modalität. Die untenstehenden Preisfelder geben, soweit verfügbar, den regulären Pay-as-you-go-Starttarif für Global/US beziehungsweise International an; Staffel-, Cache-, Batch- und Modalitätsabweichungen stehen in price_notes. Berücksichtigt sind Qwen-Text- und multimodale Chat-/Understanding-Modelle, nicht jedoch reine Generierungs-, Speech-, Embedding-, Moderations-, Realtime- oder Robotik-Modelle.", "pricing_url": "https://www.alibabacloud.com/help/en/model-studio/model-pricing", "docs_url": "https://www.alibabacloud.com/help/en/model-studio/models", "researched_at": "2026-09-04T10:48:19+0200", "usage": { "input_tokens": 111323, "output_tokens": 22446, "web_searches": 14 }, "model_count": 57, "live_checked": false, "live_error": null } ], "models": [ { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.8-Max", "api_id": "qwen3.8-max", "tier": "flagship", "status": "ga", "release_date": null, "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 2, "cached_input_price": null, "output_price": 6, "price_notes": "Internationaler Standardtarif. Context Caching wird für Eingabetokens rabattiert; in Beijing ist Batch-Datei-Eingabe günstiger.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Aktuelles multimodales Flaggschiff für anspruchsvolle Agenten-, Coding-, Dokument- und visuelle Reasoning-Aufgaben.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen3-8-max", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.8-2.4T-A95B", "api_id": "qwen3.8-2.4t-a95b", "tier": "flagship", "status": "ga", "release_date": null, "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 2, "cached_input_price": null, "output_price": 6, "price_notes": "Regulärer Preis der Qwen3.8-Max-Klasse; Context Caching wird rabattiert.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Direkt adressierbare großskalige Qwen3.8-MoE-Variante für Flaggschiff-Reasoning und multimodale Agentenaufgaben.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen-api-via-openai-responses", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.5-Omni-Plus", "api_id": "qwen3.5-omni-plus", "tier": "flagship", "status": "ga", "release_date": null, "context_window": 64000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.4, "cached_input_price": null, "output_price": 8.3, "price_notes": "Internationaler Standard für Text-/Bild-/Video-Input und Textausgabe. Audio-Input kostet 11 USD; bei Audio-only-Ausgabe beträgt der Outputpreis 44 USD pro Million Tokens.", "modalities_input": [ "text", "image", "audio", "video" ], "modalities_output": [ "text", "audio" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": false, "license": null, "description": "Aktuelles Omni-Flaggschiff für gemeinsame Text-, Bild-, Audio- und Videoanalyse mit Text- oder Sprachausgabe.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/omni", "https://www.alibabacloud.com/help/en/model-studio/qwen-omni", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.8-27B", "api_id": "qwen3.8-27b", "tier": "standard", "status": "ga", "release_date": null, "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 0.5, "cached_input_price": 0.1, "output_price": 3, "price_notes": "Internationaler Standardtarif. In Beijing beträgt der Standardtarif 0,424 USD Input und 1,696 USD Output; explizites und implizites Caching ist verfügbar.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Dichtes natives Vision-Language-Modell für Coding, Büroaufgaben und zuverlässige multimodale Ausführung.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen3-8-27b", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.7-Plus", "api_id": "qwen3.7-plus", "tier": "standard", "status": "ga", "release_date": null, "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 0.276, "cached_input_price": null, "output_price": 1.101, "price_notes": "Globaler/US-Standardtarif für Eingaben bis 256K; bei 256K bis 1M steigen Input/Output auf 0,826/3,301 USD. Thinking und Non-Thinking haben denselben Outputpreis; zeitlich begrenzte Tages-/Nachtrabatte sowie Cache-Rabatte können gelten.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Ausgewogenes multimodales Agentenmodell mit 1M Kontext für Chat, Coding, Produktivität und lange Dokumente.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen3-7-plus-us", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.5-Omni-Flash", "api_id": "qwen3.5-omni-flash", "tier": "small", "status": "ga", "release_date": null, "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.4, "cached_input_price": null, "output_price": 2.2, "price_notes": "Internationaler Standard für Text-/Bild-/Video-Input und Textausgabe. Audio-Input kostet 3 USD; bei Audio-only-Ausgabe beträgt der Outputpreis 11,9 USD pro Million Tokens.", "modalities_input": [ "text", "image", "audio", "video" ], "modalities_output": [ "text", "audio" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": false, "license": null, "description": "Schnellere und günstigere aktuelle Omni-Variante für multimodale Analyse und Sprach- oder Textantworten.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/omni", "https://www.alibabacloud.com/help/en/model-studio/qwen-omni", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.8-Flash", "api_id": "qwen3.8-flash", "tier": "small", "status": "ga", "release_date": null, "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 0.15, "cached_input_price": null, "output_price": 0.47, "price_notes": "Internationaler Standardtarif; Context Caching reduziert Eingabekosten. Die Preise in Beijing sind niedriger und Batch-Inferenz ist dort rabattiert.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Preisgünstiges multimodales 1M-Kontext-Modell für schnelle Chat-, Agenten- und Dokumentaufgaben.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen3-8-flash", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.7-Flash", "api_id": "qwen3.7-flash", "tier": "small", "status": "ga", "release_date": null, "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 0.03, "cached_input_price": null, "output_price": 0.13, "price_notes": "Internationaler Tarif für Eingaben bis 32K. Bei 32K bis 256K gelten 0,10/0,40 USD, bei 256K bis 1M 0,20/0,80 USD; Batch ist um 50 % rabattiert und Caching reduziert Eingabekosten.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Kosteneffizientes multimodales 1M-Kontext-Modell für schnelle agentische und visuelle Anwendungen.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen3-7-flash", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.5-OCR", "api_id": "qwen3.5-ocr", "tier": "specialized", "status": "ga", "release_date": "2026-06-16", "context_window": 65536, "max_output_tokens": 16384, "knowledge_cutoff": null, "input_price": 0.069, "cached_input_price": null, "output_price": 0.275, "price_notes": "Öffentlich dokumentierter Tarif für China (Beijing); die Modellseite weist keine internationalen Tokenpreise aus.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": false, "license": null, "description": "Aktuelles spezialisiertes OCR- und Dokumentverständnismodell für Extraktion, Textlokalisierung und mehrturnige Bilddialoge.", "source_urls": [ "https://docs.modelstudio.console.alibabacloud.com/en/model-studio/qwen3-5-ocr", "https://www.alibabacloud.com/help/en/model-studio/qwen-vl-ocr-api-reference", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.6-Max Preview", "api_id": "qwen3.6-max-preview", "tier": "flagship", "status": "preview", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.3, "cached_input_price": null, "output_price": 7.8, "price_notes": "Internationaler Tarif bis 128K Input; bei 128K bis 256K steigen Input/Output auf 2/12 USD. Context Caching wird rabattiert.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Noch verfügbare Vorschau des früheren Qwen3.6-Flaggschiffs mit umschaltbarem Thinking-Modus.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.7-Max Preview", "api_id": "qwen3.7-max-preview", "tier": "reasoning", "status": "preview", "release_date": null, "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 2.5, "cached_input_price": null, "output_price": 7.5, "price_notes": "Internationaler Thinking-only-Preview-Tarif; diese ID entspricht dem Snapshot vom 2026-05-17.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "streaming" ], "open_weights": false, "license": null, "description": "Vorschauversion des Qwen3.7-Max für reines textbasiertes Thinking und allgemeine Konversationsaufgaben.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen3-7-max", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.7-Max", "api_id": "qwen3.7-max", "tier": "flagship", "status": "legacy", "release_date": null, "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 2.5, "cached_input_price": null, "output_price": 7.5, "price_notes": "Internationaler Standardtarif; die undatierte ID zeigt laut Dokumentation auf qwen3.7-max-2026-05-20. Context Caching reduziert Eingabekosten.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Noch buchbares früheres Qwen3.7-Flaggschiff mit textbasierter Agenten-, Coding- und Reasoning-Ausrichtung.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen3-7-max", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-235B-A22B", "api_id": "qwen3-235b-a22b", "tier": "flagship", "status": "legacy", "release_date": "2025-05-12", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.287, "cached_input_price": null, "output_price": 1.147, "price_notes": "Globaler/US-Non-Thinking-Tarif; im Thinking-Modus bleibt der Input bei 0,287 USD, der Output beträgt 2,868 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-MoE-Modell der Qwen3-Generation mit umschaltbarem Thinking für starke allgemeine Aufgaben.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-235B-A22B-Instruct", "api_id": "qwen3-vl-235b-a22b-instruct", "tier": "flagship", "status": "legacy", "release_date": null, "context_window": 131072, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.287, "cached_input_price": null, "output_price": 1.147, "price_notes": "Globaler/US-Non-Thinking-Tarif.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Instruct-Flaggschiff für leistungsstarkes multimodales Verstehen ohne Thinking-only-Betrieb.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "QVQ-Max", "api_id": "qvq-max", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.2, "cached_input_price": null, "output_price": 4.8, "price_notes": "Internationaler Standardtarif für multimodale Tokenabrechnung.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "streaming" ], "open_weights": false, "license": null, "description": "Noch verfügbares multimodales Vision-Reasoning-Modell für anspruchsvolle Bildverständnisaufgaben.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/vision-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "QwQ-Plus", "api_id": "qwq-plus", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.8, "cached_input_price": null, "output_price": 2.4, "price_notes": "Internationaler Standardtarif; in Beijing beträgt er 0,230/0,574 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "streaming" ], "open_weights": false, "license": null, "description": "Noch buchbares spezialisiertes Text-Reasoning-Modell der QwQ-Reihe.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-235B-A22B-Thinking", "api_id": "qwen3-vl-235b-a22b-thinking", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 131072, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.287, "cached_input_price": null, "output_price": 2.868, "price_notes": "Globaler/US-Thinking-only-Tarif. In Singapore beträgt der internationale Preis 0,4/4 USD.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Vision-Language-Flaggschiff für multimodales STEM-, Mathematik- und visuelles Reasoning.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://docs.modelstudio.console.alibabacloud.com/en/model-studio/qwen3-vl-235b-a22b-thinking", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "QVQ-Plus", "api_id": "qvq-plus", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.287, "cached_input_price": null, "output_price": 0.717, "price_notes": "Öffentlich dokumentierter Tarif für China (Beijing); kein internationaler Standardtarif ist auf der aktuellen Preisseite ausgewiesen.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "streaming" ], "open_weights": false, "license": null, "description": "Kostengünstigere noch verfügbare QVQ-Variante für visuelles Reasoning.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/vision-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-235B-A22B-Thinking-2507", "api_id": "qwen3-235b-a22b-thinking-2507", "tier": "reasoning", "status": "legacy", "release_date": "2025-07", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.23, "cached_input_price": null, "output_price": 2.3, "price_notes": "Internationaler Thinking-only-Tarif. Global/US beträgt 0,23 USD Input und 2,3 USD Output.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Reasoning-Variante des 235B-A22B-Modells, bei der Thinking nicht deaktiviert werden kann.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/deep-thinking", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-32B-Thinking", "api_id": "qwen3-vl-32b-thinking", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 131072, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.16, "cached_input_price": null, "output_price": 0.64, "price_notes": "Globaler/US-Thinking-only-Tarif.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-32B-Vision-Language-Modell für ausschließlich multimodales Reasoning.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-Next-80B-A3B-Thinking", "api_id": "qwen3-next-80b-a3b-thinking", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.144, "cached_input_price": null, "output_price": 1.434, "price_notes": "Globaler/US-Thinking-only-Tarif; International beträgt er 0,15/1,2 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-MoE-Modell mittlerer Größe für ausschließlich schrittweises Reasoning.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-30B-A3B-Thinking-2507", "api_id": "qwen3-30b-a3b-thinking-2507", "tier": "reasoning", "status": "legacy", "release_date": "2025-07", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.108, "cached_input_price": null, "output_price": 1.076, "price_notes": "Globaler/US-Thinking-only-Tarif; International beträgt er 0,20/2,4 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Thinking-only-Version des 30B-A3B-MoE-Modells.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/deep-thinking", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-30B-A3B-Thinking", "api_id": "qwen3-vl-30b-a3b-thinking", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 131072, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.108, "cached_input_price": null, "output_price": 1.076, "price_notes": "Globaler/US-Thinking-only-Tarif.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-MoE-Vision-Language-Modell für effizientes multimodales Reasoning.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-8B-Thinking", "api_id": "qwen3-vl-8b-thinking", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 131072, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.072, "cached_input_price": null, "output_price": 0.717, "price_notes": "Globaler/US-Thinking-only-Tarif; International beträgt er 0,18/2,1 USD.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Kompaktes Open-Weights-Vision-Language-Modell für multimodales Thinking.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.6-27B", "api_id": "qwen3.6-27b", "tier": "standard", "status": "legacy", "release_date": null, "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.6, "cached_input_price": null, "output_price": 3.6, "price_notes": "Internationaler Standardtarif laut Preistabelle; in Beijing 0,412564/2,475384 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": false, "license": null, "description": "Noch angebotene dichte Qwen3.6-Variante für textbasierte allgemeine und Reasoning-Aufgaben.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/model-pricing", "https://www.alibabacloud.com/help/en/model-studio/text-generation-model" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.6-35B-A3B", "api_id": "qwen3.6-35b-a3b", "tier": "standard", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 0.375, "cached_input_price": null, "output_price": 2.25, "price_notes": "Internationaler Tarif bis 256K. Global/US liegt bei 0,248 USD Input und 1,485 USD Output.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "streaming" ], "open_weights": false, "license": null, "description": "Vorherige multimodale MoE-Variante mittlerer Größe für visuelle und textbasierte Agentenaufgaben.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen3-6-35b-a3b", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.6-Plus", "api_id": "qwen3.6-plus", "tier": "standard", "status": "legacy", "release_date": null, "context_window": 1000000, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 0.276, "cached_input_price": null, "output_price": 1.651, "price_notes": "Globaler/US-Tarif bis 256K Input; oberhalb davon bis 1M gelten 1,101/6,602 USD. Batch-Dateien sind rabattiert; Caching-Optionen unterscheiden sich nach Region.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Letzte vorherige Plus-Generation für multimodale Chat-, Agenten- und Langkontextaufgaben.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen3-6-plus", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-235B-A22B-Instruct-2507", "api_id": "qwen3-235b-a22b-instruct-2507", "tier": "standard", "status": "legacy", "release_date": "2025-07", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.23, "cached_input_price": null, "output_price": 0.92, "price_notes": "Internationaler Non-Thinking-Tarif; Global/US entspricht 0,23/0,92 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Instruct-Variante des 235B-A22B-Modells für direkte Antworten ohne Thinking-Modus.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/newly-released-models", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-32B", "api_id": "qwen3-32b", "tier": "standard", "status": "legacy", "release_date": "2025-05-12", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.16, "cached_input_price": null, "output_price": 0.64, "price_notes": "Globaler/US-Non-Thinking-Tarif. Im Thinking-Modus ist der Outputpreis ebenfalls 0,64 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Dense-Modell mittlerer Größe mit hybridem Thinking für allgemeine Aufgaben.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/newly-released-models", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-32B-Instruct", "api_id": "qwen3-vl-32b-instruct", "tier": "standard", "status": "legacy", "release_date": null, "context_window": 131072, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.16, "cached_input_price": null, "output_price": 0.64, "price_notes": "Globaler/US-Non-Thinking-Tarif.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-32B-Vision-Language-Instruct-Modell für Bild-, Video- und Textverständnis.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-Next-80B-A3B-Instruct", "api_id": "qwen3-next-80b-a3b-instruct", "tier": "standard", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.144, "cached_input_price": null, "output_price": 0.574, "price_notes": "Globaler/US-Non-Thinking-Tarif; International beträgt er 0,15/1,2 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Instruct-Variante des kompakten Qwen3-Next-MoE-Modells.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-14B", "api_id": "qwen3-14b", "tier": "standard", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.144, "cached_input_price": null, "output_price": 0.574, "price_notes": "Globaler/US-Non-Thinking-Tarif; Thinking-Output kostet 1,434 USD pro Million Tokens.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Dense-Modell für kompaktere allgemeine Chat- und Reasoning-Workloads.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-Plus", "api_id": "qwen3-vl-plus", "tier": "standard", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.143, "cached_input_price": null, "output_price": 1.434, "price_notes": "Globaler/US-Tarif bis 32K Input; bei 32K bis 128K gelten 0,215/2,15 USD, bei 128K bis 256K 0,43/4,301 USD. Context Caching reduziert Eingabekosten.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Noch buchbare vorherige kommerzielle Vision-Language-Plus-Variante mit hybridem Thinking.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/vision-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-30B-A3B", "api_id": "qwen3-30b-a3b", "tier": "standard", "status": "legacy", "release_date": "2025-05-12", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.108, "cached_input_price": null, "output_price": 0.431, "price_notes": "Globaler/US-Non-Thinking-Tarif; im Thinking-Modus beträgt der Outputpreis 1,076 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-MoE-Modell mit hybridem Thinking für gute Effizienz bei allgemeinen Aufgaben.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/newly-released-models", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-30B-A3B-Instruct-2507", "api_id": "qwen3-30b-a3b-instruct-2507", "tier": "standard", "status": "legacy", "release_date": "2025-07", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.108, "cached_input_price": null, "output_price": 0.431, "price_notes": "Globaler/US-Non-Thinking-Tarif; International beträgt er 0,20/0,80 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Instruct-Version des 30B-A3B-MoE-Modells für direkte Antworten.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-30B-A3B-Instruct", "api_id": "qwen3-vl-30b-a3b-instruct", "tier": "standard", "status": "legacy", "release_date": "2025-10-03", "context_window": 131072, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.108, "cached_input_price": null, "output_price": 0.431, "price_notes": "Globaler/US-Non-Thinking-Tarif.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-MoE-Instruct-Modell für lange multimodale Dokument- und Videoaufgaben.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://docs.modelstudio.console.alibabacloud.com/en/model-studio/newly-released-models", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-Coder-480B-A35B-Instruct", "api_id": "qwen3-coder-480b-a35b-instruct", "tier": "coding", "status": "legacy", "release_date": "2025-07-23", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.861, "cached_input_price": null, "output_price": 3.441, "price_notes": "Globaler/US-Tarif bis 32K Input; bei 32K bis 128K gelten 1,291/5,161 USD, bei 128K bis 200K 2,151/8,602 USD. Der internationale Preis ist höher.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Coding-MoE-Modell mit hoher Coding-Agent-Leistung für große Softwareaufgaben.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/newly-released-models", "https://www.alibabacloud.com/help/en/model-studio/model-pricing", "https://www.alibabacloud.com/help/en/model-studio/model-telemetry" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-Coder-Plus", "api_id": "qwen3-coder-plus", "tier": "coding", "status": "legacy", "release_date": null, "context_window": 1000000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.574, "cached_input_price": null, "output_price": 2.294, "price_notes": "Globaler/US-Tarif bis 32K Input; bei 32K bis 128K gelten 0,861/3,441 USD, bei 128K bis 256K 1,434/5,735 USD und bei 256K bis 1M 2,868/28,671 USD. Context Caching reduziert Eingabekosten.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Vorheriges leistungsstarkes Coding-Agent-Modell mit 1M Kontext und Tool Calling.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-Coder-Next", "api_id": "qwen3-coder-next", "tier": "coding", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.3, "cached_input_price": null, "output_price": 1.5, "price_notes": "Internationaler Tarif bis 32K Input; weitere Stufen sind 0,5/2,5 USD bis 128K und 0,8/4 USD bis 256K. In Beijing gelten niedrigere Preise.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": false, "license": null, "description": "Kompaktes vorheriges Coding-Modell für agentisches Programmieren und Tool Calling.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-Coder-30B-A3B-Instruct", "api_id": "qwen3-coder-30b-a3b-instruct", "tier": "coding", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.216, "cached_input_price": null, "output_price": 0.861, "price_notes": "Globaler/US-Tarif bis 32K Input; bei 32K bis 128K gelten 0,323/1,291 USD und bei 128K bis 200K 0,538/2,151 USD.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Kompakteres Open-Weights-Coding-MoE-Modell für Codegenerierung und Tool-gestützte Entwicklung.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-Coder-Flash", "api_id": "qwen3-coder-flash", "tier": "coding", "status": "legacy", "release_date": null, "context_window": 1000000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.144, "cached_input_price": null, "output_price": 0.574, "price_notes": "Globaler/US-Tarif bis 32K Input; weitere Stufen sind 0,216/0,861 USD bis 128K, 0,359/1,434 USD bis 256K und 0,717/3,584 USD bis 1M. Context Caching reduziert Eingabekosten.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Schnellere und preiswertere vorherige Coding-Variante mit 1M Kontext.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-Omni-Flash", "api_id": "qwen3-omni-flash", "tier": "small", "status": "legacy", "release_date": null, "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.43, "cached_input_price": null, "output_price": 1.66, "price_notes": "Internationaler Text-only-Standard. Der Preis variiert nach Text-, Audio- und Bild/Video-Input sowie Text- oder Audioausgabe; bei multimodalem Input beträgt der Textoutput 3,06 USD und Audiooutput 15,11 USD.", "modalities_input": [ "text", "image", "audio", "video" ], "modalities_output": [ "text", "audio" ], "capabilities": [ "reasoning", "tool_use", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Noch buchbare vorherige leichte Omni-Generation mit Thinking und multimodalem Ein- und Ausgabeumfang.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/omni", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3.6-Flash", "api_id": "qwen3.6-flash", "tier": "small", "status": "legacy", "release_date": null, "context_window": 1000000, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 0.25, "cached_input_price": null, "output_price": 1.5, "price_notes": "Internationaler Tarif bis 256K Input; bei 256K bis 1M gelten 1/4 USD. Batch-Inferenz ist um 50 % rabattiert und Context Caching reduziert Eingabekosten.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Letzte vorherige schnelle multimodale Qwen3.6-Variante für preisbewusste Anwendungen.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-8B", "api_id": "qwen3-8b", "tier": "small", "status": "legacy", "release_date": "2025-05-12", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.072, "cached_input_price": null, "output_price": 0.287, "price_notes": "Globaler/US-Non-Thinking-Tarif; Thinking-Output kostet 0,717 USD. Für einige Aufrufarten ist Thinking laut Dokumentation nur mit Streaming nutzbar.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Kleinmodell mit hybridem Thinking für latenz- und kostenbewusste Textanwendungen.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/newly-released-models", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-8B-Instruct", "api_id": "qwen3-vl-8b-instruct", "tier": "small", "status": "legacy", "release_date": null, "context_window": 131072, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.072, "cached_input_price": null, "output_price": 0.287, "price_notes": "Globaler/US-Non-Thinking-Tarif; International beträgt er 0,18/0,70 USD.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming", "fine_tuning" ], "open_weights": true, "license": "Apache-2.0", "description": "Kompaktes Open-Weights-Vision-Language-Instruct-Modell für lange Videos, Dokumente und räumliche Aufgaben.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://docs.modelstudio.console.alibabacloud.com/en/model-studio/qwen3-vl-8b-instruct", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-Flash", "api_id": "qwen3-vl-flash", "tier": "small", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.022, "cached_input_price": null, "output_price": 0.215, "price_notes": "Globaler/US-Tarif bis 32K Input; bei 32K bis 128K gelten 0,043/0,43 USD und bei 128K bis 256K 0,086/0,859 USD. Context Caching reduziert Eingabekosten.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Schnelle vorherige kommerzielle Vision-Language-Variante für kostengünstiges Bild- und Videoverständnis.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/vision-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-4B", "api_id": "qwen3-4b", "tier": "small", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Für die direkt nutzbare Modell-ID nennt die aktuelle öffentliche Pay-as-you-go-Preisseite keinen separaten Tokenpreis.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Open-Weights-Kleinmodell der Qwen3-Reihe für lokale oder kostensensitive allgemeine Textaufgaben.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/text-generation-model" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-1.7B", "api_id": "qwen3-1.7b", "tier": "small", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Für die direkt nutzbare Modell-ID nennt die aktuelle öffentliche Pay-as-you-go-Preisseite keinen separaten Tokenpreis.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Sehr kleines Open-Weights-Qwen3-Modell für ressourcensparende Text- und Chat-Workloads.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/text-generation-model" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-0.6B", "api_id": "qwen3-0.6b", "tier": "small", "status": "legacy", "release_date": null, "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Für die direkt nutzbare Modell-ID nennt die aktuelle öffentliche Pay-as-you-go-Preisseite keinen separaten Tokenpreis.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Kleinstes Open-Weights-Modell der Qwen3-Reihe für sehr ressourcenarme Textaufgaben.", "source_urls": [ "https://github.com/QwenLM/Qwen3", "https://www.alibabacloud.com/help/en/model-studio/text-generation-model" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-4B-Instruct", "api_id": "qwen3-vl-4b-instruct", "tier": "small", "status": "legacy", "release_date": null, "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Für diese Modell-ID weist die aktuelle öffentliche Pay-as-you-go-Preisseite keinen separaten Tokenpreis aus; sie ist als Modell für Bereitstellung/Fine-Tuning dokumentiert.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming", "fine_tuning" ], "open_weights": true, "license": "Apache-2.0", "description": "Kleines Open-Weights-Vision-Language-Instruct-Modell für ressourcenbewusstes multimodales Verstehen.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://www.alibabacloud.com/help/en/model-studio/model-training-and-deployment-billing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen3-VL-2B-Instruct", "api_id": "qwen3-vl-2b-instruct", "tier": "small", "status": "legacy", "release_date": null, "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Für diese Modell-ID weist die aktuelle öffentliche Pay-as-you-go-Preisseite keinen separaten Tokenpreis aus; sie ist als Modell für Bereitstellung dokumentiert.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Sehr kleines Open-Weights-Vision-Language-Instruct-Modell für einfache multimodale Anwendungen.", "source_urls": [ "https://github.com/QwenLM/Qwen3-VL", "https://www.alibabacloud.com/help/en/model-studio/model-training-and-deployment-billing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen-Math-Plus", "api_id": "qwen-math-plus", "tier": "specialized", "status": "legacy", "release_date": null, "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.574, "cached_input_price": null, "output_price": 1.721, "price_notes": "Öffentlich dokumentierter Tarif für China (Beijing); qwen-math-plus-latest wird nicht als separates Modell gezählt.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "streaming" ], "open_weights": false, "license": null, "description": "Spezialisiertes noch buchbares Textmodell für mathematische Aufgaben und schrittweises Problemlösen.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen-Plus-Character-JA", "api_id": "qwen-plus-character-ja", "tier": "specialized", "status": "legacy", "release_date": null, "context_window": 32000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.5, "cached_input_price": null, "output_price": 1.4, "price_notes": "Internationaler, in Singapore ausgewiesener Standardtarif; Session Cache wird rabattiert.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": false, "license": null, "description": "Japanisch ausgerichtete Variante des Charakter- und Rollenspiel-Chatmodells.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen-Plus-Character", "api_id": "qwen-plus-character", "tier": "specialized", "status": "legacy", "release_date": null, "context_window": 32000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.115, "cached_input_price": null, "output_price": 0.287, "price_notes": "Globaler/US-Standardtarif; international in Singapore 0,5/1,4 USD. Session Cache wird rabattiert.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": false, "license": null, "description": "Spezialisiertes Chatmodell für Rollen-, Charakter- und dialogorientierte Interaktionen.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen-Long", "api_id": "qwen-long", "tier": "specialized", "status": "legacy", "release_date": null, "context_window": 10000000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.072, "cached_input_price": null, "output_price": 0.287, "price_notes": "Öffentlich dokumentierter Tarif für China (Beijing). qwen-long-latest verweist auf die aktuelle Version und wird hier nicht als separates Modell gezählt.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "structured_output", "streaming" ], "open_weights": false, "license": null, "description": "Spezialmodell mit 10M Kontext für die Verarbeitung sehr langer Dokumente und umfangreicher Textsammlungen.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen-VL-OCR", "api_id": "qwen-vl-ocr", "tier": "specialized", "status": "legacy", "release_date": null, "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.043, "cached_input_price": null, "output_price": 0.072, "price_notes": "Globaler/US-Standardtarif; die undatierte ID entspricht laut Preisseite dem Snapshot vom 2025-11-20.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": false, "license": null, "description": "Noch buchbares vorheriges spezialisiertes OCR-Modell für Dokumente, Tabellen und Text aus Bildern.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/qwen-vl-ocr-api-reference", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "alibaba", "provider": "Alibaba (Qwen)", "name": "Qwen-Flash-Character", "api_id": "qwen-flash-character", "tier": "specialized", "status": "legacy", "release_date": null, "context_window": 8000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.034, "cached_input_price": null, "output_price": 0.203, "price_notes": "Globaler/US-Standardtarif; international in Singapore 0,05/0,4 USD. Session Cache wird rabattiert.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": false, "license": null, "description": "Schnelle, kostengünstige Charakter-Chatvariante für kurze dialogorientierte Interaktionen.", "source_urls": [ "https://www.alibabacloud.com/help/en/model-studio/text-generation-model", "https://www.alibabacloud.com/help/en/model-studio/model-pricing" ], "verified_in_api": null }, { "provider_id": "amazon", "provider": "Amazon (Nova)", "name": "Amazon Nova 2 Lite", "api_id": "amazon.nova-2-lite-v1:0", "tier": "standard", "status": "ga", "release_date": "2025-12-02", "context_window": 1000000, "max_output_tokens": 65536, "knowledge_cutoff": "2025-10", "input_price": 0.3, "cached_input_price": 0.075, "output_price": 2.5, "price_notes": "Standardtarif für Text-Tokens. Batch-Inferenz wird laut Bedrock-Preisseite mit 50 % Rabatt gegenüber On-Demand berechnet; Flex und Priority sind als weitere Service-Tiers verfügbar. Reasoning-Tokens zählen als Output-Tokens.", "modalities_input": [ "text", "image", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "web_search", "code_execution", "fine_tuning", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Kosteneffizientes multimodales Reasoning-Modell für Automatisierung, Dokumentenverarbeitung, Kundenservice und agentische Aufgaben.", "source_urls": [ "https://aws.amazon.com/bedrock/pricing/", "https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-amazon-nova-2-lite.html", "https://aws.amazon.com/about-aws/whats-new/2025/12/nova-2-foundation-models-amazon-bedrock/", "https://docs.aws.amazon.com/nova/latest/nova2-userguide/whats-new.html" ], "verified_in_api": null }, { "provider_id": "amazon", "provider": "Amazon (Nova)", "name": "Amazon Nova 2 Pro", "api_id": "amazon.nova-pro-v2:0", "tier": "flagship", "status": "preview", "release_date": "2025-12-02", "context_window": 1000000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Für die zugangsbeschränkte Preview veröffentlicht AWS keinen allgemein verfügbaren Token-Standardtarif auf der Bedrock-Preisseite. Der Early Access steht Nova-Forge-Kunden über den AWS-Account-Support offen.", "modalities_input": [ "text", "image", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "fine_tuning", "streaming" ], "open_weights": false, "license": null, "description": "Intelligentestes Nova-Modell für komplexe mehrstufige Reasoning-, Planungs-, Videoanalyse- und Softwaremigrationsaufgaben.", "source_urls": [ "https://aws.amazon.com/about-aws/whats-new/2025/12/nova-2-foundation-models-amazon-bedrock/", "https://aws.amazon.com/nova/models/", "https://docs.aws.amazon.com/general/latest/gr/bedrock.html", "https://aws.amazon.com/bedrock/pricing/" ], "verified_in_api": null }, { "provider_id": "amazon", "provider": "Amazon (Nova)", "name": "Amazon Nova 2 Omni", "api_id": "amazon.nova-2-omni-v1:0", "tier": "specialized", "status": "preview", "release_date": "2025-12-02", "context_window": 1000000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.3, "cached_input_price": 0.075, "output_price": 2.5, "price_notes": "Standardtarif für Text-Eingabe und Text-Ausgabe; die Bildausgabe wird zusätzlich in einer separaten, nicht tokenbasierten Preisdimension geführt und ist daher nicht in output_price enthalten. Der Zugriff ist als Nova-Forge-Early-Access-Preview eingeschränkt.", "modalities_input": [ "text", "image", "audio", "video" ], "modalities_output": [ "text", "image" ], "capabilities": [ "reasoning", "tool_use", "streaming" ], "open_weights": false, "license": null, "description": "All-in-one-Previewmodell für multimodales Reasoning mit Text-, Bild-, Video- und Spracheingaben sowie Text- und Bildausgaben.", "source_urls": [ "https://aws.amazon.com/about-aws/whats-new/2025/12/amazon-nova-2-omni-preview/", "https://docs.aws.amazon.com/nova/latest/nova2-userguide/request-response-schema.html", "https://aws.amazon.com/bedrock/pricing/", "https://builder.aws.com/content/3B2n4BYrroKX8KFkn1rF7Y2LegQ/brio-building-a-unified-creator-workspace-with-amazon-nova-on-amazon-bedrock" ], "verified_in_api": null }, { "provider_id": "amazon", "provider": "Amazon (Nova)", "name": "Amazon Nova Premier", "api_id": "amazon.nova-premier-v1:0", "tier": "flagship", "status": "legacy", "release_date": "2025-10-31", "context_window": 1000000, "max_output_tokens": 25000, "knowledge_cutoff": "2024-10", "input_price": 2.5, "cached_input_price": 0.25, "output_price": 12.5, "price_notes": "Standardtarif für On-Demand-Tokeninferenz. Batch-Inferenz ist mit 50 % Rabatt gegenüber On-Demand verfügbar; Flex und Priority werden ebenfalls unterstützt. AWS kennzeichnet dieses Modell als Legacy und nennt als EOL-Datum den 14. September 2026.", "modalities_input": [ "text", "image", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Legacy-Flaggschiff für komplexes multimodales Reasoning, agentische Workflows und Modelldestillation.", "source_urls": [ "https://aws.amazon.com/bedrock/pricing/", "https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-amazon-nova-premier.html", "https://docs.aws.amazon.com/nova/latest/userguide/what-is-nova.html" ], "verified_in_api": null }, { "provider_id": "amazon", "provider": "Amazon (Nova)", "name": "Amazon Nova Pro", "api_id": "amazon.nova-pro-v1:0", "tier": "standard", "status": "legacy", "release_date": "2024-12-05", "context_window": 300000, "max_output_tokens": 5000, "knowledge_cutoff": "2024-10", "input_price": 0.8, "cached_input_price": 0.08, "output_price": 3.2, "price_notes": "Standardtarif für On-Demand-Tokeninferenz. Batch-Inferenz kostet 50 % weniger als On-Demand; Flex und Priority werden unterstützt. Prompt-Cache-Lesezugriffe erhalten gegenüber regulären Eingabetokens einen Rabatt.", "modalities_input": [ "text", "image", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "fine_tuning", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Legacy-Allroundmodell mit ausgewogenem Verhältnis von Qualität, Geschwindigkeit und Kosten für multimodale und Code-Aufgaben.", "source_urls": [ "https://aws.amazon.com/bedrock/pricing/", "https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-amazon-nova-pro.html", "https://docs.aws.amazon.com/nova/latest/userguide/what-is-nova.html" ], "verified_in_api": null }, { "provider_id": "amazon", "provider": "Amazon (Nova)", "name": "Amazon Nova Lite", "api_id": "amazon.nova-lite-v1:0", "tier": "small", "status": "legacy", "release_date": "2024-12-05", "context_window": 300000, "max_output_tokens": 5000, "knowledge_cutoff": "2024-10", "input_price": 0.06, "cached_input_price": 0.006, "output_price": 0.24, "price_notes": "Standardtarif für On-Demand-Tokeninferenz. Batch-Inferenz kostet 50 % weniger als On-Demand. Prompt-Cache-Lesezugriffe erhalten gegenüber regulären Eingabetokens einen Rabatt.", "modalities_input": [ "text", "image", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "fine_tuning", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Legacy-Kleinmodell für schnelle und günstige multimodale Verarbeitung von Text, Bildern und Video.", "source_urls": [ "https://aws.amazon.com/bedrock/pricing/", "https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-amazon-nova-lite.html", "https://docs.aws.amazon.com/nova/latest/userguide/what-is-nova.html" ], "verified_in_api": null }, { "provider_id": "amazon", "provider": "Amazon (Nova)", "name": "Amazon Nova Micro", "api_id": "amazon.nova-micro-v1:0", "tier": "small", "status": "legacy", "release_date": "2024-12-05", "context_window": 128000, "max_output_tokens": 5000, "knowledge_cutoff": "2024-10", "input_price": 0.035, "cached_input_price": 0.0035, "output_price": 0.14, "price_notes": "Standardtarif für On-Demand-Tokeninferenz. Batch-Inferenz kostet 50 % weniger als On-Demand. Prompt-Cache-Lesezugriffe erhalten gegenüber regulären Eingabetokens einen Rabatt.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "fine_tuning", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Legacy-Textmodell mit besonders niedriger Latenz und Kosten für Klassifikation, Extraktion, Übersetzung und Zusammenfassungen.", "source_urls": [ "https://aws.amazon.com/bedrock/pricing/", "https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-amazon-nova-micro.html", "https://docs.aws.amazon.com/nova/latest/userguide/what-is-nova.html" ], "verified_in_api": null }, { "provider_id": "anthropic", "provider": "Anthropic", "name": "Claude Fable 5.1", "api_id": "claude-fable-5-1", "tier": "flagship", "status": "ga", "release_date": "2026-09-01", "context_window": 1000000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-06", "input_price": 10, "cached_input_price": 0.25, "output_price": 50, "price_notes": "Prompt-Cache-Schreiben: 12,50 USD/MTok (5 Min.) bzw. 20 USD/MTok (1 Std.); Cache-Hits kosten 0,25 USD/MTok. Die Batch API reduziert Ein- und Ausgabetokens um 50 %; bei US-Inferenzgeografie gilt auf unterstützten Plattformen ein 1,1-facher Aufschlag.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Anthropics leistungsstärkstes breit verfügbares Modell für anspruchsvolles Reasoning sowie lang laufende agentische Coding-, Recherche- und Wissensarbeit.", "source_urls": [ "https://platform.claude.com/docs/en/models/fable-5-1/overview", "https://platform.claude.com/docs/en/about-claude/pricing", "https://platform.claude.com/docs/en/build-with-claude/structured-outputs" ], "verified_in_api": true }, { "provider_id": "anthropic", "provider": "Anthropic", "name": "Claude Sonnet 5", "api_id": "claude-sonnet-5", "tier": "standard", "status": "ga", "release_date": "2026-06-30", "context_window": 1000000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-01", "input_price": 2, "cached_input_price": 0.2, "output_price": 10, "price_notes": "Prompt-Cache-Schreiben: 2,50 USD/MTok (5 Min.) bzw. 4 USD/MTok (1 Std.); Cache-Hits kosten 0,20 USD/MTok. Die Batch API reduziert Ein- und Ausgabetokens um 50 %; die zum 1. September 2026 geplante Preiserhöhung auf 3/15 USD/MTok wurde nicht umgesetzt.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Ausgewogenes Standardmodell für schnelle, intelligente Coding-, Agenten-, Analyse- und Unternehmensworkloads.", "source_urls": [ "https://platform.claude.com/docs/en/models/sonnet-5/overview", "https://platform.claude.com/docs/en/about-claude/pricing", "https://platform.claude.com/docs/en/build-with-claude/structured-outputs" ], "verified_in_api": true }, { "provider_id": "anthropic", "provider": "Anthropic", "name": "Claude Opus 5", "api_id": "claude-opus-5", "tier": "coding", "status": "ga", "release_date": "2026-07-24", "context_window": 1000000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-05", "input_price": 5, "cached_input_price": 0.5, "output_price": 25, "price_notes": "Prompt-Cache-Schreiben: 6,25 USD/MTok (5 Min.) bzw. 10 USD/MTok (1 Std.); Cache-Hits kosten 0,50 USD/MTok. Batch API bietet 50 % Rabatt; Fast Mode ist als Research Preview auf der First-Party-API verfügbar und kostet 10 USD Input bzw. 50 USD Output pro MTok; im Batch-Modus ist Fast Mode nicht verfügbar.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Leistungsstarkes Opus-Modell für komplexes agentisches Coding, Enterprise-Aufgaben und tiefes adaptives Reasoning.", "source_urls": [ "https://platform.claude.com/docs/en/models/opus-5/overview", "https://platform.claude.com/docs/en/about-claude/pricing", "https://platform.claude.com/docs/en/build-with-claude/fast-mode" ], "verified_in_api": true }, { "provider_id": "anthropic", "provider": "Anthropic", "name": "Claude Haiku 4.5", "api_id": "claude-haiku-4-5", "tier": "small", "status": "ga", "release_date": "2025-10-15", "context_window": 200000, "max_output_tokens": 64000, "knowledge_cutoff": "2025-02", "input_price": 1, "cached_input_price": 0.1, "output_price": 5, "price_notes": "Die undatierte API-ID ist ein Alias für den aktuellen festen Snapshot. Prompt-Cache-Schreiben: 1,25 USD/MTok (5 Min.) bzw. 2 USD/MTok (1 Std.); Cache-Hits kosten 0,10 USD/MTok; Batch API bietet 50 % Rabatt.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Kleines, besonders schnelles und kostengünstiges Modell für hohe Volumina mit manuell steuerbarem Extended Thinking.", "source_urls": [ "https://platform.claude.com/docs/en/models/haiku-4-5/overview", "https://platform.claude.com/docs/en/about-claude/models/model-ids-and-versions", "https://platform.claude.com/docs/en/about-claude/pricing" ], "verified_in_api": true }, { "provider_id": "anthropic", "provider": "Anthropic", "name": "Claude Mythos 5.1", "api_id": "claude-mythos-5-1", "tier": "specialized", "status": "ga", "release_date": "2026-09-01", "context_window": 1000000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-06", "input_price": 10, "cached_input_price": 0.25, "output_price": 50, "price_notes": "Nur auf Einladung für Project-Glasswing-Teilnehmende verfügbar. Prompt-Cache-Schreiben: 12,50 USD/MTok (5 Min.) bzw. 20 USD/MTok (1 Std.); Cache-Hits kosten 0,25 USD/MTok; Batch API mit 50 % Rabatt auf Input und Output.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Zugangsbeschränkte Project-Glasswing-Variante von Fable 5.1 mit gleichen Spezifikationen für spezialisierte Sicherheitsforschung.", "source_urls": [ "https://platform.claude.com/docs/en/models/mythos-5-1/overview", "https://platform.claude.com/docs/en/about-claude/pricing", "https://platform.claude.com/docs/en/build-with-claude/structured-outputs" ], "verified_in_api": false }, { "provider_id": "anthropic", "provider": "Anthropic", "name": "Claude Mythos 5", "api_id": "claude-mythos-5", "tier": "specialized", "status": "ga", "release_date": "2026-06-09", "context_window": 1000000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-01", "input_price": 10, "cached_input_price": 1, "output_price": 50, "price_notes": "Nur auf Einladung im Rahmen von Project Glasswing verfügbar und inzwischen durch Mythos 5.1 ergänzt. Prompt-Cache-Schreiben: 12,50 USD/MTok (5 Min.) bzw. 20 USD/MTok (1 Std.); Cache-Hits kosten 1 USD/MTok; Batch API bietet 50 % Rabatt.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Zugangsbeschränktes Spezialmodell für defensive Cybersicherheits- und Biologieforschung ohne die Safety-Classifier von Fable 5.", "source_urls": [ "https://platform.claude.com/docs/en/models/mythos-5/overview", "https://platform.claude.com/docs/en/about-claude/pricing", "https://platform.claude.com/docs/en/build-with-claude/structured-outputs" ], "verified_in_api": false }, { "provider_id": "anthropic", "provider": "Anthropic", "name": "Claude Fable 5", "api_id": "claude-fable-5", "tier": "flagship", "status": "legacy", "release_date": "2026-06-09", "context_window": 1000000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-01", "input_price": 10, "cached_input_price": 1, "output_price": 50, "price_notes": "Weiter verfügbar, aber von Fable 5.1 abgelöst. Prompt-Cache-Schreiben: 12,50 USD/MTok (5 Min.) bzw. 20 USD/MTok (1 Std.); Cache-Hits kosten 1 USD/MTok; Batch API bietet 50 % Rabatt.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Älteres Flaggschiff für lang laufende agentische Aufgaben und anspruchsvolles Reasoning mit stets aktiviertem adaptivem Thinking.", "source_urls": [ "https://platform.claude.com/docs/en/models/fable-5/overview", "https://platform.claude.com/docs/en/about-claude/pricing", "https://platform.claude.com/docs/en/build-with-claude/structured-outputs" ], "verified_in_api": true }, { "provider_id": "anthropic", "provider": "Anthropic", "name": "Claude Sonnet 4.6", "api_id": "claude-sonnet-4-6", "tier": "standard", "status": "legacy", "release_date": "2026-02-17", "context_window": 1000000, "max_output_tokens": 128000, "knowledge_cutoff": "2025-08", "input_price": 3, "cached_input_price": 0.3, "output_price": 15, "price_notes": "Weiter verfügbar, aber von Sonnet 5 abgelöst. Prompt-Cache-Schreiben: 3,75 USD/MTok (5 Min.) bzw. 6 USD/MTok (1 Std.); Cache-Hits kosten 0,30 USD/MTok; Batch API bietet 50 % Rabatt.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Unmittelbarer Sonnet-Vorgänger für ausgewogene Unternehmens-, Coding- und Agentenaufgaben mit adaptivem Thinking.", "source_urls": [ "https://platform.claude.com/docs/en/models/sonnet-4-6/overview", "https://platform.claude.com/docs/en/about-claude/pricing", "https://platform.claude.com/docs/en/build-with-claude/structured-outputs" ], "verified_in_api": true }, { "provider_id": "anthropic", "provider": "Anthropic", "name": "Claude Opus 4.8", "api_id": "claude-opus-4-8", "tier": "coding", "status": "legacy", "release_date": "2026-05-28", "context_window": 1000000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-01", "input_price": 5, "cached_input_price": 0.5, "output_price": 25, "price_notes": "Weiter verfügbar, aber von Opus 5 abgelöst. Prompt-Cache-Schreiben: 6,25 USD/MTok (5 Min.) bzw. 10 USD/MTok (1 Std.); Cache-Hits kosten 0,50 USD/MTok; Batch API bietet 50 % Rabatt. Fast Mode ist als Research Preview auf der First-Party-API für 10 USD Input bzw. 50 USD Output pro MTok verfügbar und nicht mit Batch kombinierbar.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Unmittelbarer Opus-Vorgänger für anspruchsvolles agentisches Coding und adaptives Reasoning.", "source_urls": [ "https://platform.claude.com/docs/en/models/opus-4-8/overview", "https://platform.claude.com/docs/en/about-claude/pricing", "https://platform.claude.com/docs/en/build-with-claude/fast-mode" ], "verified_in_api": true }, { "provider_id": "cohere", "provider": "Cohere", "name": "Command A", "api_id": "command-a", "tier": "flagship", "status": "ga", "release_date": "2025-03-13", "context_window": 256000, "max_output_tokens": 8000, "knowledge_cutoff": "2024-06-01", "input_price": 2.5, "cached_input_price": null, "output_price": 10, "price_notes": "Regulärer Standardtarif pro 1 Mio. Tokens.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "CC-BY-NC mit Acceptable-Use-Addendum", "description": "Leistungsstarkes Enterprise-Modell für Tool-Nutzung, RAG, Agenten und mehrsprachige Textaufgaben.", "source_urls": [ "https://docs.cohere.com/docs/command-a", "https://cohere.com/blog/command-a", "https://cohere.com/research/papers/command-a-technical-report.pdf" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Command A+", "api_id": "command-a-plus", "tier": "flagship", "status": "ga", "release_date": "2026-05-20", "context_window": 128000, "max_output_tokens": 64000, "knowledge_cutoff": "2025-04-01", "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein regulärer Token-Standardtarif veröffentlicht. Laut Dokumentation für Trial- und Production-Keys bis zum Rate-Limit kostenlos; produktiver Einsatz über Model Vault.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Cohere-Flaggschiff als MoE-Modell für multimodale, mehrsprachige und reasoning-intensiven Agenten-Workloads.", "source_urls": [ "https://docs.cohere.com/docs/command-a-plus", "https://docs.cohere.com/v2/changelog", "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Command A Reasoning", "api_id": "command-a-reasoning", "tier": "reasoning", "status": "ga", "release_date": "2025-08-21", "context_window": 256000, "max_output_tokens": 32000, "knowledge_cutoff": "2024-06-01", "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein regulärer Token-Standardtarif veröffentlicht. Laut Dokumentation bis zum Rate-Limit kostenlos; produktive Nutzung über Model Vault.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": false, "license": null, "description": "Hybrid-Reasoning-Modell für komplexe, mehrstufige Agenten- und Problemlösungsaufgaben.", "source_urls": [ "https://docs.cohere.com/docs/command-a-reasoning", "https://docs.cohere.com/v2/changelog" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Aya Expanse 32B", "api_id": "c4ai-aya-expanse-32b", "tier": "standard", "status": "ga", "release_date": "2024-10-24", "context_window": 128000, "max_output_tokens": 4000, "knowledge_cutoff": null, "input_price": 0.5, "cached_input_price": null, "output_price": 1.5, "price_notes": "Der offizielle Pricing-FAQ nennt diesen Tarif für Aya Expanse 8B und 32B; das 8B-API-Modell wurde am 4. April 2026 eingestellt.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Mehrsprachiges 32B-Modell für Textgenerierung, Zusammenfassung und Übersetzung in 23 Sprachen.", "source_urls": [ "https://docs.cohere.com/docs/aya-expanse", "https://cohere.com/pricing", "https://cohere.com/models-overview", "https://cohere.com/blog/tag/research" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Aya Vision 32B", "api_id": "c4ai-aya-vision-32b", "tier": "standard", "status": "ga", "release_date": "2025-03-04", "context_window": 16000, "max_output_tokens": 4000, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein Token-Standardtarif für Aya Vision 32B in den offiziellen Preisangaben belegt.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Mehrsprachiges multimodales 32B-Modell für Bildbeschreibung, visuelle Fragen, Textgenerierung und Übersetzung.", "source_urls": [ "https://docs.cohere.com/docs/aya-vision", "https://cohere.com/models-overview", "https://cohere.com/blog/tag/research" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "North Mini Code", "api_id": "north-mini-code-1-0", "tier": "coding", "status": "ga", "release_date": "2026-06-09", "context_window": 256000, "max_output_tokens": 64000, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein regulärer Token-Standardtarif veröffentlicht. Bis zum Rate-Limit kostenlos; produktive Nutzung über Model Vault.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Effizientes MoE-Coding-Modell für agentisches Programmieren und terminalbasierte Software-Engineering-Aufgaben.", "source_urls": [ "https://docs.cohere.com/docs/north-mini-code-1.0", "https://docs.cohere.com/v2/changelog", "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Command R7B", "api_id": "command-r7b", "tier": "small", "status": "ga", "release_date": "2024-12-13", "context_window": 128000, "max_output_tokens": 4000, "knowledge_cutoff": "2024-06-01", "input_price": 0.0375, "cached_input_price": null, "output_price": 0.15, "price_notes": "Regulärer Standardtarif pro 1 Mio. Tokens.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Kleines und schnelles Enterprise-Modell für kostensensitive RAG-, Tool-Use- und Agenten-Anwendungen.", "source_urls": [ "https://docs.cohere.com/docs/command-r7b", "https://docs.cohere.com/changelog/command-r-7b/", "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Tiny Aya Base", "api_id": null, "tier": "small", "status": "ga", "release_date": "2026-02-17", "context_window": 8000, "max_output_tokens": 8000, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Pretrained Open-Weights-Basismodell ohne Cohere-Chat-API-Zugang; daher kein API-Tokenpreis.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Vortrainiertes kompaktes 3,35B-Basismodell für mehrsprachige Anwendungen in 70 Sprachen.", "source_urls": [ "https://docs.cohere.com/docs/tiny-aya", "https://cohere.com/blog/cohere-labs-tiny-aya" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Tiny Aya Global", "api_id": "tiny-aya-global", "tier": "small", "status": "ga", "release_date": "2026-02-17", "context_window": 8000, "max_output_tokens": 8000, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein Token-Standardtarif in den offiziellen Cohere-Preisangaben belegt.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Kompaktes instruktionsgetuntes 3,35B-Modell mit ausgewogener Leistung über Regionen und 70 Sprachen.", "source_urls": [ "https://docs.cohere.com/docs/tiny-aya", "https://docs.cohere.com/docs/models", "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Tiny Aya Earth", "api_id": "tiny-aya-earth", "tier": "small", "status": "ga", "release_date": "2026-02-17", "context_window": 8000, "max_output_tokens": 8000, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein Token-Standardtarif in den offiziellen Cohere-Preisangaben belegt.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Kompaktes instruktionsgetuntes 3,35B-Modell, besonders auf westasiatische und afrikanische Sprachen ausgerichtet.", "source_urls": [ "https://docs.cohere.com/docs/tiny-aya", "https://docs.cohere.com/docs/models", "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Tiny Aya Fire", "api_id": "tiny-aya-fire", "tier": "small", "status": "ga", "release_date": "2026-02-17", "context_window": 8000, "max_output_tokens": 8000, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein Token-Standardtarif in den offiziellen Cohere-Preisangaben belegt.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Kompaktes instruktionsgetuntes 3,35B-Modell, besonders auf südasiatische Sprachen ausgerichtet.", "source_urls": [ "https://docs.cohere.com/docs/tiny-aya", "https://docs.cohere.com/docs/models", "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Tiny Aya Water", "api_id": "tiny-aya-water", "tier": "small", "status": "ga", "release_date": "2026-02-17", "context_window": 8000, "max_output_tokens": 8000, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein Token-Standardtarif in den offiziellen Cohere-Preisangaben belegt.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Kompaktes instruktionsgetuntes 3,35B-Modell, besonders auf europäische und asiatisch-pazifische Sprachen ausgerichtet.", "source_urls": [ "https://docs.cohere.com/docs/tiny-aya", "https://docs.cohere.com/docs/models", "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Command A Translate", "api_id": "command-a-translate", "tier": "specialized", "status": "ga", "release_date": "2025-08-28", "context_window": 8000, "max_output_tokens": 8000, "knowledge_cutoff": "2024-06-01", "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein regulärer Token-Standardtarif veröffentlicht. Bis zum Rate-Limit kostenlos; für Produktion verweist Cohere auf den Vertrieb.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "structured_output", "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Spezialisiertes Modell für kontextbewusste maschinelle Übersetzung in 23 Sprachen.", "source_urls": [ "https://docs.cohere.com/docs/command-a-translate", "https://docs.cohere.com/v2/changelog", "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Command A Vision", "api_id": "command-a-vision", "tier": "specialized", "status": "ga", "release_date": "2025-07-31", "context_window": 128000, "max_output_tokens": 8000, "knowledge_cutoff": "2024-06-01", "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Kein regulärer Token-Standardtarif veröffentlicht. Bis zum Rate-Limit kostenlos; für Produktion verweist Cohere auf den Vertrieb.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "structured_output", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Multimodales Enterprise-Modell für Bild- und Dokumentverständnis, OCR, Diagramme und Tabellen.", "source_urls": [ "https://docs.cohere.com/docs/command-a-vision", "https://docs.cohere.com/v1/changelog/2025-07-31-command-a-vision", "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "North Micro Vision", "api_id": null, "tier": "specialized", "status": "ga", "release_date": null, "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Open-Weights-Angebot ohne in den offiziellen Quellen dokumentierten Cohere-Chat-API-Tarif oder API-Modell-ID.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [], "open_weights": true, "license": "Apache-2.0", "description": "Kompaktes Vision-Language-Modell für anpassbares Bild- und Dokumentverständnis mit nativer Bildauflösung.", "source_urls": [ "https://cohere.com/models-overview" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Command R+", "api_id": "command-r-plus", "tier": "flagship", "status": "legacy", "release_date": "2024-08-30", "context_window": 128000, "max_output_tokens": 4000, "knowledge_cutoff": "2024-06-01", "input_price": 2.5, "cached_input_price": null, "output_price": 10, "price_notes": "Regulärer Standardtarif pro 1 Mio. Tokens. Als Live-Snapshot gelistet, aber von Cohere für die meisten Fälle durch Command A abgelöst.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Älteres großes Modell für dialogorientierte Langkontext-, RAG- und mehrstufige Tool-Use-Workflows.", "source_urls": [ "https://docs.cohere.com/docs/command-r-plus", "https://docs.cohere.com/v1/changelog/command-gets-refreshed", "https://cohere.com/pricing" ], "verified_in_api": null }, { "provider_id": "cohere", "provider": "Cohere", "name": "Command R", "api_id": "command-r", "tier": "standard", "status": "legacy", "release_date": "2024-08-30", "context_window": 128000, "max_output_tokens": 4000, "knowledge_cutoff": "2024-06-01", "input_price": 0.15, "cached_input_price": null, "output_price": 0.6, "price_notes": "Regulärer Standardtarif pro 1 Mio. Tokens. Als Live-Snapshot gelistet, aber von Cohere für die meisten Fälle durch Command A abgelöst.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "streaming" ], "open_weights": true, "license": "CC-BY-NC-4.0", "description": "Älteres skalierbares Modell für dialogorientierte Langkontext-, RAG- und Tool-Use-Aufgaben.", "source_urls": [ "https://docs.cohere.com/docs/command-r", "https://docs.cohere.com/v1/changelog/command-gets-refreshed", "https://cohere.com/pricing" ], "verified_in_api": null }, { "provider_id": "deepseek", "provider": "DeepSeek", "name": "DeepSeek-V4-Pro", "api_id": "deepseek-v4-pro", "tier": "flagship", "status": "ga", "release_date": "2026-08-13", "context_window": 1000000, "max_output_tokens": 384000, "knowledge_cutoff": null, "input_price": 1.32, "cached_input_price": 0.044, "output_price": 3.96, "price_notes": "input_price, cached_input_price und output_price entsprechen dem nicht rabattierten Peak-Tarif in USD pro 1 Mio. Tokens. Außerhalb der Peak-Zeiten gelten jeweils die halbierten Off-Peak-Preise: 0,66 USD Input-Cache-Miss, 0,022 USD Cache-Hit und 1,98 USD Output.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT License", "description": "Flaggschiffmodell für anspruchsvolle Reasoning-, Agenten- und Coding-Workflows mit einstellbarem Reasoning-Aufwand bis Max.", "source_urls": [ "https://api-docs.deepseek.com/quick_start/pricing/", "https://api-docs.deepseek.com/news/news260813/", "https://api-docs.deepseek.com/guides/thinking_mode/", "https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813" ], "verified_in_api": true }, { "provider_id": "deepseek", "provider": "DeepSeek", "name": "DeepSeek-V4-Flash", "api_id": "deepseek-v4-flash", "tier": "standard", "status": "preview", "release_date": "2026-07-31", "context_window": 1000000, "max_output_tokens": 384000, "knowledge_cutoff": null, "input_price": 0.44, "cached_input_price": 0.014, "output_price": 1.32, "price_notes": "input_price, cached_input_price und output_price entsprechen dem nicht rabattierten Peak-Tarif in USD pro 1 Mio. Tokens. Außerhalb der Peak-Zeiten gelten jeweils die halbierten Off-Peak-Preise: 0,22 USD Input-Cache-Miss, 0,007 USD Cache-Hit und 0,66 USD Output.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT License", "description": "Schnelles und kosteneffizientes V4-Textmodell für allgemeine Chat-, Agenten-, Coding- und Reasoning-Aufgaben mit umschaltbarem Thinking-Modus.", "source_urls": [ "https://api-docs.deepseek.com/quick_start/pricing/", "https://api-docs.deepseek.com/news/news260424/", "https://api-docs.deepseek.com/updates/", "https://api-docs.deepseek.com/guides/thinking_mode/", "https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731" ], "verified_in_api": true }, { "provider_id": "deepseek", "provider": "DeepSeek", "name": "DeepSeek-V4-Flash-Vision-Exp", "api_id": "deepseek-v4-flash-vision-exp", "tier": "specialized", "status": "preview", "release_date": "2026-08-21", "context_window": 1000000, "max_output_tokens": 384000, "knowledge_cutoff": null, "input_price": 0.44, "cached_input_price": 0.014, "output_price": 1.32, "price_notes": "input_price, cached_input_price und output_price entsprechen dem nicht rabattierten Peak-Tarif in USD pro 1 Mio. Tokens; außerhalb der Peak-Zeiten gelten jeweils die halbierten Off-Peak-Preise. Eingabebilder werden abhängig von ihren Abmessungen in Tokens umgerechnet und wie Input-Tokens zum V4-Flash-Tarif abgerechnet.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT License", "description": "Experimentelles multimodales V4-Flash-Modell für Text- und Bildverständnis in Chat- und Agenten-Workflows.", "source_urls": [ "https://api-docs.deepseek.com/quick_start/pricing/", "https://api-docs.deepseek.com/news/news260821/", "https://api-docs.deepseek.com/guides/vision/", "https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-Vision-Exp" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 3.8 Flash", "api_id": "gemini-3.8-flash", "tier": "flagship", "status": "ga", "release_date": "2026-09-02", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 0.75, "cached_input_price": 0.075, "output_price": 3.75, "price_notes": "Einführungspreis bis einschließlich 2026-12-31; ab 2027-01-01 1,50 USD Input und 7,50 USD Output. Batch und Flex kosten jeweils 50 % des Standardtarifs, Priority 180 %; Cache-Speicherung wird zusätzlich berechnet.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Googles leistungsstärkstes Flash-Modell für langlaufende Softwareentwicklung, autonome Agenten und komplexe Unternehmensabläufe.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/changelog" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemma 4 31B IT", "api_id": "gemma-4-31b-it", "tier": "standard", "status": "ga", "release_date": null, "context_window": 262144, "max_output_tokens": null, "knowledge_cutoff": "2025-01", "input_price": 0, "cached_input_price": null, "output_price": 0, "price_notes": "Über die Gemini API ist das gehostete Gemma-4-Angebot im Free Tier kostenlos; für den Paid Tier weist Google keinen Tokenpreis aus.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "web_search", "streaming" ], "open_weights": true, "license": "Apache 2.0", "description": "Offenes dichtes 31B-Instruct-Modell für multimodales Text- und Bildverständnis, Coding und Reasoning.", "source_urls": [ "https://ai.google.dev/gemma/docs/core/gemma_on_gemini_api", "https://ai.google.dev/gemma/docs/core/model_card_4", "https://ai.google.dev/gemini-api/docs/pricing" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 3.7 Flash", "api_id": "gemini-3.7-flash", "tier": "coding", "status": "ga", "release_date": "2026-08-13", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 0.75, "cached_input_price": 0.075, "output_price": 3.75, "price_notes": "Einführungspreis bis einschließlich 2026-12-31; ab 2027-01-01 1,50 USD Input und 7,50 USD Output. Batch und Flex kosten jeweils 50 % des Standardtarifs, Priority 180 %; Cache-Speicherung wird zusätzlich berechnet.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Schnelles multimodales Reasoning-Modell für Coding, Tool-Nutzung und zuverlässige mehrstufige Agentenabläufe.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/changelog" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 3.6 Flash", "api_id": "gemini-3.6-flash", "tier": "coding", "status": "ga", "release_date": "2026-07-21", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 0.75, "cached_input_price": 0.075, "output_price": 3.75, "price_notes": "Einführungspreis bis einschließlich 2026-12-31; ab 2027-01-01 1,50 USD Input und 7,50 USD Output. Batch und Flex kosten jeweils 50 % des Standardtarifs, Priority 180 %; Cache-Speicherung wird zusätzlich berechnet.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Schnelles Frontier-Modell für Codegenerierung, agentische Ausführung und iterative Coding-Zyklen.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/changelog" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 3.5 Flash-Lite", "api_id": "gemini-3.5-flash-lite", "tier": "small", "status": "ga", "release_date": "2026-07-21", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 0.3, "cached_input_price": 0.03, "output_price": 2.5, "price_notes": "Standardpreis gilt für Text-, Bild-, Video- und Audioeingaben. Batch und Flex: 0,15 USD Input sowie 1,25 USD Output; Priority: 0,54 USD Input sowie 4,50 USD Output. Cache-Speicherung wird zusätzlich berechnet.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Kleines, multimodales Low-Latency-Modell für kostengünstige Hochdurchsatz- und Subagentenaufgaben.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/changelog" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemma 4 26B A4B IT", "api_id": "gemma-4-26b-a4b-it", "tier": "small", "status": "ga", "release_date": null, "context_window": 262144, "max_output_tokens": null, "knowledge_cutoff": "2025-01", "input_price": 0, "cached_input_price": null, "output_price": 0, "price_notes": "Über die Gemini API ist das gehostete Gemma-4-Angebot im Free Tier kostenlos; für den Paid Tier weist Google keinen Tokenpreis aus.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "web_search", "streaming" ], "open_weights": true, "license": "Apache 2.0", "description": "Offenes 26B-MoE-Instruct-Modell mit 3,8B aktiven Parametern für effizientes multimodales Reasoning und Coding.", "source_urls": [ "https://ai.google.dev/gemma/docs/core/gemma_on_gemini_api", "https://ai.google.dev/gemma/docs/core/model_card_4", "https://ai.google.dev/gemini-api/docs/pricing" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 3.1 Pro Preview", "api_id": "gemini-3.1-pro-preview", "tier": "flagship", "status": "preview", "release_date": "2026-02-19", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 2, "cached_input_price": 0.2, "output_price": 12, "price_notes": "Standardtarif für Prompts bis 200.000 Tokens; oberhalb davon 4,00 USD Input, 0,40 USD Cache-Input und 18,00 USD Output. Batch und Flex kosten bei bis zu 200.000 Tokens 1,00 USD Input und 6,00 USD Output; Priority kostet 3,60 USD Input und 21,60 USD Output.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Preview-Pro-Modell für präzises multimodales Reasoning, Software Engineering und zuverlässige Tool-gestützte Agentenabläufe.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/deprecations" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 3 Flash Preview", "api_id": "gemini-3-flash-preview", "tier": "reasoning", "status": "preview", "release_date": "2025-12-17", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 0.5, "cached_input_price": 0.05, "output_price": 3, "price_notes": "Standardpreis gilt für Text-, Bild- und Videoeingaben; Audio kostet 1,00 USD Input und 0,10 USD für gecachten Input. Batch und Flex: 0,25 USD Input für Text/Bild/Video sowie 1,50 USD Output; Priority: 0,90 USD Input für Text/Bild/Video und 5,40 USD Output.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Älteres Preview-Modell für multimodales Verständnis sowie agentische und Coding-orientierte Aufgaben.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-3-flash-preview", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/deprecations" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 3.1 Pro Preview Custom Tools", "api_id": "gemini-3.1-pro-preview-customtools", "tier": "coding", "status": "preview", "release_date": "2026-02-19", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 2, "cached_input_price": 0.2, "output_price": 12, "price_notes": "Separater API-Endpunkt derselben Pro-Preview-Familie mit identischem Tarif; optimiert für die Priorisierung eigener Werkzeuge. Für Prompts über 200.000 Tokens gelten 4,00 USD Input und 18,00 USD Output; Batch, Flex und Priority weichen wie bei Gemini 3.1 Pro Preview ab.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Spezialisierter Pro-Preview-Endpunkt für Coding-Agenten, die Bash- und eigene Funktionswerkzeuge priorisieren sollen.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview", "https://ai.google.dev/gemini-api/docs/pricing" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini Deep Research", "api_id": "deep-research-preview", "tier": "specialized", "status": "preview", "release_date": "2026-04-21", "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Google berechnet die tatsächliche Nutzung anhand der zugrunde liegenden Gemini-Inferenz einschließlich Zwischen- und Reasoning-Tokens sowie verwendeter Tools; kein fester modellbezogener Tokenpreis ist ausgewiesen.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "web_search", "code_execution", "streaming" ], "open_weights": false, "license": null, "description": "Preview-Forschungsagent für effiziente mehrstufige Web-Recherche und das Streamen von Ergebnissen an Client-Oberflächen.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/deep-research", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/changelog" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini Deep Research Max", "api_id": "deep-research-max-preview", "tier": "specialized", "status": "preview", "release_date": "2026-04-21", "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Google berechnet die tatsächliche Nutzung anhand der zugrunde liegenden Gemini-Inferenz einschließlich Zwischen- und Reasoning-Tokens sowie verwendeter Tools; kein fester modellbezogener Tokenpreis ist ausgewiesen.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "web_search", "code_execution", "streaming" ], "open_weights": false, "license": null, "description": "Preview-Max-Variante des Forschungsagenten für besonders umfassende Kontextsammlung, Wettbewerbsanalysen und Due Diligence.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/deep-research", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/changelog" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Antigravity Agent", "api_id": "antigravity-preview", "tier": "specialized", "status": "preview", "release_date": "2026-05-19", "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Abrechnung nach den Tokenpreisen des gewählten zugrunde liegenden Gemini-Modells und der Tool-Nutzung; während der Preview berechnet Google keine Sandbox-Compute-Ressourcen. Standardmäßig verwendet der Agent Gemini 3.8 Flash, das Modell ist aber konfigurierbar.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "web_search", "code_execution", "streaming" ], "open_weights": false, "license": null, "description": "Preview-Managed-Agent mit Google-gehosteter Linux-Sandbox für Planung, Codeausführung, Dateiverwaltung und Web-Browsing.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/antigravity-agent", "https://ai.google.dev/gemini-api/docs/agents", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/changelog" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 2.5 Pro", "api_id": "gemini-2.5-pro", "tier": "reasoning", "status": "legacy", "release_date": "2025-06-17", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": "2025-01", "input_price": 1.25, "cached_input_price": 0.125, "output_price": 10, "price_notes": "Weiterhin verfügbares Vorgänger-Pro-Modell. Standardtarif gilt bis 200.000 Prompt-Tokens; darüber 2,50 USD Input, 0,25 USD Cache-Input und 15,00 USD Output. Batch und Flex: 0,625/5,00 USD bei bis zu 200.000 Tokens; Priority: 2,25/18,00 USD.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Älteres Pro-Reasoning-Modell für komplexe Code-, Mathematik-, STEM- und Long-Context-Aufgaben.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-2.5-pro", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/deprecations" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 3.5 Flash", "api_id": "gemini-3.5-flash", "tier": "standard", "status": "legacy", "release_date": "2026-05-19", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 1.5, "cached_input_price": 0.15, "output_price": 9, "price_notes": "Weiterhin buchbares, in der Modellübersicht als Legacy bezeichnetes Flash-Modell. Batch und Flex: 0,75 USD Input sowie 4,50 USD Output; Priority: 2,70 USD Input sowie 16,20 USD Output. Cache-Speicherung wird zusätzlich berechnet.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Vorgänger-Flash-Modell für skalierbare Subagenten, mehrstufige Workflows und langlaufende Coding-Aufgaben.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/changelog" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 2.5 Flash", "api_id": "gemini-2.5-flash", "tier": "standard", "status": "legacy", "release_date": "2025-06-17", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": "2025-01", "input_price": 0.3, "cached_input_price": 0.03, "output_price": 2.5, "price_notes": "Weiterhin verfügbares Vorgänger-Flash-Modell. Standardpreis gilt für Text-, Bild- und Videoeingaben; Audio kostet 1,00 USD Input und 0,10 USD für gecachten Input. Batch und Flex: 0,15 USD Input für Text/Bild/Video sowie 1,25 USD Output; Priority: 0,54 USD Input und 4,50 USD Output.", "modalities_input": [ "text", "image", "audio", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Älteres preisleistungsstarkes multimodales Hybrid-Reasoning-Modell für hohe Volumina und Agenten.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-2.5-flash", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/deprecations" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 3.1 Flash-Lite", "api_id": "gemini-3.1-flash-lite", "tier": "small", "status": "legacy", "release_date": "2026-05-07", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": null, "input_price": 0.25, "cached_input_price": 0.025, "output_price": 1.5, "price_notes": "Standardpreis gilt für Text-, Bild- und Videoeingaben; Audio kostet 0,50 USD Input und 0,05 USD für gecachten Input. Batch und Flex halbieren die Tokenpreise; Priority kostet 0,45 USD Input für Text/Bild/Video bzw. 0,90 USD für Audio sowie 2,70 USD Output. Google nennt als frühestes Abschaltdatum 2027-05-07 und empfiehlt Gemini 3.5 Flash-Lite.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Älteres kleines multimodales Modell für einfache, sehr häufige und latenzkritische Aufgaben.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-3.1-flash-lite", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/deprecations" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 2.5 Flash-Lite", "api_id": "gemini-2.5-flash-lite", "tier": "small", "status": "legacy", "release_date": "2025-07-22", "context_window": 1048576, "max_output_tokens": 65536, "knowledge_cutoff": "2025-01", "input_price": 0.1, "cached_input_price": 0.01, "output_price": 0.4, "price_notes": "Weiterhin verfügbares Vorgänger-Kleinmodell. Standardpreis gilt für Text-, Bild- und Videoeingaben; Audio kostet 0,30 USD Input und 0,03 USD für gecachten Input. Batch und Flex halbieren die Tokenpreise; Priority ist teurer.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Älteres sehr kostengünstiges multimodales Modell für Klassifikation, Extraktion und extrem niedrige Latenz.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-2.5-flash-lite", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/deprecations" ], "verified_in_api": true }, { "provider_id": "google", "provider": "Google (Gemini)", "name": "Gemini 2.5 Computer Use Preview", "api_id": "gemini-2.5-computer-use-preview", "tier": "specialized", "status": "legacy", "release_date": "2025-10-07", "context_window": null, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.25, "cached_input_price": null, "output_price": 10, "price_notes": "Älteres spezialisiertes Computer-Use-Modell. Tarif bis 200.000 Prompt-Tokens; oberhalb davon 2,50 USD Input und 15,00 USD Output. Google empfiehlt inzwischen Computer Use mit Gemini-3.x-Modellen.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "computer_use", "streaming" ], "open_weights": false, "license": null, "description": "Legacy-Modell zur Steuerung von Browser-, Mobil- und Desktop-Oberflächen anhand von Screenshots und generierten UI-Aktionen.", "source_urls": [ "https://ai.google.dev/gemini-api/docs/computer-use", "https://ai.google.dev/gemini-api/docs/pricing", "https://ai.google.dev/gemini-api/docs/changelog" ], "verified_in_api": true }, { "provider_id": "meta", "provider": "Meta (Llama)", "name": "Llama 4 Maverick", "api_id": null, "tier": "flagship", "status": "ga", "release_date": "2025-04-05", "context_window": 1000000, "max_output_tokens": null, "knowledge_cutoff": "2024-08", "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Open Weights ohne direkte Meta-Token-API; Meta veröffentlicht hierfür keinen eigenen API-Standardtarif.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "fine_tuning" ], "open_weights": true, "license": "Llama 4 Community License", "description": "Das leistungsstärkere offene Llama-4-Multimodalmodell für Bild- und Textverständnis sowie allgemeine Chat- und Agentenaufgaben.", "source_urls": [ "https://ai.meta.com/llama/get-started/", "https://github.com/meta-llama/llama-models/blob/main/models/llama4/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama4/LICENSE" ], "verified_in_api": null }, { "provider_id": "meta", "provider": "Meta (Llama)", "name": "Llama 4 Scout", "api_id": null, "tier": "standard", "status": "ga", "release_date": "2025-04-05", "context_window": 10000000, "max_output_tokens": null, "knowledge_cutoff": "2024-08", "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Open Weights ohne direkte Meta-Token-API; Meta veröffentlicht hierfür keinen eigenen API-Standardtarif.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "fine_tuning" ], "open_weights": true, "license": "Llama 4 Community License", "description": "Ein effizientes offenes Llama-4-Multimodalmodell mit sehr großem Kontextfenster, das auf den Betrieb mit einer einzelnen H100 ausgelegt ist.", "source_urls": [ "https://ai.meta.com/llama/get-started/", "https://github.com/meta-llama/llama-models/blob/main/models/llama4/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama4/LICENSE" ], "verified_in_api": null }, { "provider_id": "meta", "provider": "Meta (Llama)", "name": "Muse Glimmer 30B", "api_id": null, "tier": "small", "status": "ga", "release_date": "2026-08-10", "context_window": 131072, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Keine direkte, von Meta betriebene Token-API für dieses Open-Weights-Modell; die Kosten hängen bei Selbstbetrieb von der eigenen Infrastruktur ab.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "fine_tuning" ], "open_weights": true, "license": "Apache License 2.0", "description": "Ein lokal ausführbares, 30B großes multimodales Agentenmodell für mehrstufige Tool- und Automatisierungsaufgaben auf eigener Hardware.", "source_urls": [ "https://huggingface.co/meta-models/Muse-Glimmer-30B", "https://github.com/meta-models/meta-oss-cookbook", "https://ai.meta.com/ai-for-good/" ], "verified_in_api": null }, { "provider_id": "meta", "provider": "Meta (Llama)", "name": "Muse Spark 1.3", "api_id": "muse-spark-1.3", "tier": "flagship", "status": "preview", "release_date": "2026-09-02", "context_window": 1048576, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 1.25, "cached_input_price": 0.15, "output_price": 4.25, "price_notes": "Standardtarif. Die alternative API-ID „muse-spark-1.3-contributor“ bezeichnet denselben Checkpoint, kostet 0,10 USD Input, 0,002 USD Cached Input und 0,20 USD Output pro 1 Mio. Tokens; dabei dürfen Eingaben und Ausgaben zur Verbesserung von Meta-Produkten verwendet werden. Reasoning-Tokens werden zum Output-Tarif abgerechnet.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "prompt_caching" ], "open_weights": false, "license": null, "description": "Metas aktuelles multimodales Flaggschiff für langlaufende agentische Aufgaben, Tool-Nutzung und Coding-Workflows.", "source_urls": [ "https://research.meta.ai/blog/introducing-muse-spark-1-3", "https://dev.meta.ai/docs/models" ], "verified_in_api": null }, { "provider_id": "meta", "provider": "Meta (Llama)", "name": "Llama 3.3 70B Instruct", "api_id": null, "tier": "standard", "status": "legacy", "release_date": "2024-12-06", "context_window": 131072, "max_output_tokens": null, "knowledge_cutoff": "2023-12", "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Open Weights ohne direkte Meta-Token-API; Meta veröffentlicht hierfür keinen eigenen API-Standardtarif.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "fine_tuning" ], "open_weights": true, "license": "Llama 3.3 Community License", "description": "Der weiterhin verfügbare Text-only-Vorgänger für mehrsprachige Assistenz-, Dialog- und Code-Workloads.", "source_urls": [ "https://github.com/meta-llama/llama-models/blob/main/models/llama3_3/MODEL_CARD.md", "https://github.com/meta-llama/llama-models/blob/main/models/llama3_3/LICENSE" ], "verified_in_api": null }, { "provider_id": "meta", "provider": "Meta (Llama)", "name": "Muse Spark 1.2", "api_id": "muse-spark-1.2", "tier": "coding", "status": "legacy", "release_date": "2026-08-05", "context_window": 1048576, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 1.25, "cached_input_price": 0.15, "output_price": 4.25, "price_notes": "Standardtarif. Die alternative API-ID „muse-spark-1.2-contributor“ ist derselbe Checkpoint mit 0,10 USD Input, 0,002 USD Cached Input und 0,20 USD Output pro 1 Mio. Tokens; Metas Produktverbesserungsnutzung der Anfrage- und Antwortdaten ist dafür zulässig. Reasoning-Tokens gelten als Output-Tokens.", "modalities_input": [ "text", "image", "audio", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "prompt_caching" ], "open_weights": false, "license": null, "description": "Weiterhin buchbarer Vorgänger von Muse Spark 1.3, der besonders für Codegenerierung, komplexes Debugging und große Repository-Workflows optimiert wurde.", "source_urls": [ "https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2", "https://dev.meta.ai/docs/models" ], "verified_in_api": null }, { "provider_id": "minimax", "provider": "MiniMax", "name": "MiniMax M3", "api_id": "MiniMax-M3", "tier": "flagship", "status": "ga", "release_date": "2026-06-01", "context_window": 1000000, "max_output_tokens": 524288, "knowledge_cutoff": null, "input_price": 0.3, "cached_input_price": 0.06, "output_price": 1.2, "price_notes": "Standardtarif für bis zu 512k Input-Tokens. Bei mehr als 512k Input-Tokens: 0,60 USD Input, 0,12 USD Cache-Read und 2,40 USD Output pro 1 Mio. Tokens. Priority-Service kostet jeweils 1,5× des Standardtarifs.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "prompt_caching", "streaming" ], "open_weights": true, "license": "MiniMax Community License", "description": "MiniMax’ multimodales Flaggschiff für agentisches Reasoning, Tool-Nutzung, Coding sowie sehr lange Kontexte bis 1 Mio. Tokens.", "source_urls": [ "https://platform.minimax.io/docs/guides/pricing-paygo", "https://platform.minimax.io/docs/guides/text-generation", "https://platform.minimax.io/docs/api-reference/text-chat-openai", "https://platform.minimax.io/docs/release-notes/models", "https://platform.minimax.io/docs/guides/local-deploy-m3" ], "verified_in_api": null }, { "provider_id": "minimax", "provider": "MiniMax", "name": "MiniMax M2.7 Highspeed", "api_id": "MiniMax-M2.7-highspeed", "tier": "coding", "status": "ga", "release_date": "2026-03-18", "context_window": 204800, "max_output_tokens": 204800, "knowledge_cutoff": null, "input_price": 0.6, "cached_input_price": 0.06, "output_price": 2.4, "price_notes": "Schnellere Variante mit laut MiniMax gleicher Modellleistung; Prompt-Cache-Write kostet 0,375 USD pro 1 Mio. Tokens.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "prompt_caching", "streaming" ], "open_weights": true, "license": "NON-COMMERCIAL LICENSE", "description": "Die auf niedrigere Latenz und etwa 100 Tokens pro Sekunde ausgelegte Highspeed-Variante von MiniMax M2.7.", "source_urls": [ "https://platform.minimax.io/docs/guides/pricing-paygo", "https://platform.minimax.io/docs/guides/text-generation", "https://platform.minimax.io/docs/release-notes/models", "https://www.minimax.io/models/text/m27", "https://huggingface.co/MiniMaxAI/MiniMax-M2.7/blob/main/LICENSE" ], "verified_in_api": null }, { "provider_id": "minimax", "provider": "MiniMax", "name": "MiniMax M2.7", "api_id": "MiniMax-M2.7", "tier": "coding", "status": "ga", "release_date": "2026-03-18", "context_window": 204800, "max_output_tokens": 204800, "knowledge_cutoff": null, "input_price": 0.3, "cached_input_price": 0.06, "output_price": 1.2, "price_notes": "Prompt-Cache-Write kostet 0,375 USD pro 1 Mio. Tokens.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "prompt_caching", "streaming" ], "open_weights": true, "license": "NON-COMMERCIAL LICENSE", "description": "Ein auf Software Engineering, komplexe Agentenaufgaben, Tool-Nutzung und produktive Wissensarbeit ausgerichtetes Reasoning- und Coding-Modell.", "source_urls": [ "https://platform.minimax.io/docs/guides/pricing-paygo", "https://platform.minimax.io/docs/guides/text-generation", "https://platform.minimax.io/docs/release-notes/models", "https://platform.minimax.io/docs/guides/local-deploy-m2-7", "https://huggingface.co/MiniMaxAI/MiniMax-M2.7/blob/main/LICENSE" ], "verified_in_api": null }, { "provider_id": "minimax", "provider": "MiniMax", "name": "MiniMax M2.5 Highspeed", "api_id": "MiniMax-M2.5-highspeed", "tier": "coding", "status": "legacy", "release_date": "2026-02", "context_window": 204800, "max_output_tokens": 204800, "knowledge_cutoff": null, "input_price": 0.6, "cached_input_price": 0.03, "output_price": 2.4, "price_notes": "Als Legacy-Modell weiterhin buchbar; schnellere Variante mit gleicher Modellleistung. Prompt-Cache-Write kostet 0,375 USD pro 1 Mio. Tokens.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "prompt_caching", "streaming" ], "open_weights": true, "license": "MiniMax Model License", "description": "Die schnellere, weiterhin verfügbare Legacy-Variante von M2.5 für Coding- und Agent-Aufgaben.", "source_urls": [ "https://platform.minimax.io/docs/guides/pricing-paygo", "https://platform.minimax.io/docs/guides/text-generation", "https://platform.minimax.io/docs/release-notes/models", "https://www.minimax.io/models/text", "https://huggingface.co/MiniMaxAI/MiniMax-M2.5/blob/main/LICENSE-MODEL" ], "verified_in_api": null }, { "provider_id": "minimax", "provider": "MiniMax", "name": "MiniMax M2.5", "api_id": "MiniMax-M2.5", "tier": "coding", "status": "legacy", "release_date": "2026-02", "context_window": 204800, "max_output_tokens": 204800, "knowledge_cutoff": null, "input_price": 0.3, "cached_input_price": 0.03, "output_price": 1.2, "price_notes": "Als Legacy-Modell weiterhin buchbar; Prompt-Cache-Write kostet 0,375 USD pro 1 Mio. Tokens.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "prompt_caching", "streaming" ], "open_weights": true, "license": "MiniMax Model License", "description": "Die letzte noch buchbare Vorgängergeneration für Coding, Agent-Workflows, Suche und komplexe Office-Aufgaben.", "source_urls": [ "https://platform.minimax.io/docs/guides/pricing-paygo", "https://platform.minimax.io/docs/guides/text-generation", "https://platform.minimax.io/docs/release-notes/models", "https://www.minimax.io/models/text", "https://huggingface.co/MiniMaxAI/MiniMax-M2.5/blob/main/LICENSE-MODEL" ], "verified_in_api": null }, { "provider_id": "minimax", "provider": "MiniMax", "name": "M2-her", "api_id": "M2-her", "tier": "specialized", "status": "legacy", "release_date": null, "context_window": 64000, "max_output_tokens": 2048, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "In der aktuellen Pay-as-you-go-Preistabelle ist kein separater Tarif mehr ausgewiesen; daher keine Preisangabe. Das Modell ist weiterhin per API-Aufruf mit der festen Modell-ID „M2-her“ dokumentiert.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "streaming" ], "open_weights": false, "license": null, "description": "Ein spezialisiertes Text-Chatmodell für Rollenspiel, reichhaltige Rollenprofile und lange Mehrturn-Dialoge.", "source_urls": [ "https://platform.minimax.io/docs/guides/text-chat", "https://platform.minimax.io/docs/guides/text-generation", "https://platform.minimax.io/docs/guides/pricing-paygo" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Mistral Medium 3.5", "api_id": "mistral-medium-3-5", "tier": "flagship", "status": "ga", "release_date": "2026-04-28", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.5, "cached_input_price": 0.15, "output_price": 7.5, "price_notes": "Standardtarif. Batch kostet 50 % weniger; regionale Inferenz kostet für unterstützte Modelle 10 % mehr. Cache-Hits kosten 0,15 USD pro 1 Mio. Input-Tokens.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching" ], "open_weights": true, "license": "Modified MIT", "description": "Frontier-Multimodalmodell für langlaufende agentische Aufgaben, einstellbares Reasoning und Softwareentwicklung.", "source_urls": [ "https://docs.mistral.ai/models/mistral-medium-3-5-26-04", "https://docs.mistral.ai/inference/pricing", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Mistral Large 3", "api_id": "mistral-large-3", "tier": "flagship", "status": "ga", "release_date": "2025-12-02", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.5, "cached_input_price": 0.05, "output_price": 1.5, "price_notes": "Standardtarif. Batch kostet 50 % weniger; regionale Inferenz kostet für unterstützte Modelle 10 % mehr. Cache-Hits kosten 0,05 USD pro 1 Mio. Input-Tokens.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching" ], "open_weights": true, "license": "Apache 2.0", "description": "Offenes multimodales Flaggschiff-Allzweckmodell mit granularer Mixture-of-Experts-Architektur.", "source_urls": [ "https://docs.mistral.ai/models/mistral-large-3-25-12", "https://docs.mistral.ai/inference/pricing", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Codestral", "api_id": "codestral", "tier": "coding", "status": "ga", "release_date": "2025-07-30", "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.3, "cached_input_price": 0.03, "output_price": 0.9, "price_notes": "Standardtarif. Batch kostet 50 % weniger; regionale Inferenz kostet für unterstützte Modelle 10 % mehr. Cache-Hits kosten 0,03 USD pro 1 Mio. Input-Tokens.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "batch", "prompt_caching" ], "open_weights": false, "license": null, "description": "Niedrig-latentes Coding-Modell für häufige Code-Vervollständigung, Fill-in-the-Middle und Codegenerierung.", "source_urls": [ "https://docs.mistral.ai/models/codestral-25-08", "https://docs.mistral.ai/inference/pricing", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Ministral 3 14B", "api_id": "ministral-3-14b", "tier": "small", "status": "ga", "release_date": "2025-12-02", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.2, "cached_input_price": 0.02, "output_price": 0.2, "price_notes": "Standardtarif. Batch kostet 50 % weniger; regionale Inferenz kostet für unterstützte Modelle 10 % mehr. Cache-Hits kosten 0,02 USD pro 1 Mio. Input-Tokens.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "batch", "prompt_caching" ], "open_weights": true, "license": "Apache 2.0", "description": "Größtes Modell der edge-orientierten Ministral-3-Familie für leistungsfähige lokale Text- und Vision-Anwendungen.", "source_urls": [ "https://docs.mistral.ai/models/ministral-3-14b-25-12", "https://docs.mistral.ai/inference/pricing", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Mistral Small 4", "api_id": "mistral-small-4", "tier": "small", "status": "ga", "release_date": "2026-03-16", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.15, "cached_input_price": 0.015, "output_price": 0.6, "price_notes": "Standardtarif. Batch kostet 50 % weniger; regionale Inferenz kostet für unterstützte Modelle 10 % mehr. Cache-Hits kosten 0,015 USD pro 1 Mio. Input-Tokens.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching" ], "open_weights": true, "license": "Apache 2.0", "description": "Effizientes multimodales Hybridmodell, das Instruction-Following, Reasoning und Coding in einem Modell vereint.", "source_urls": [ "https://docs.mistral.ai/models/mistral-small-4-0-26-03", "https://docs.mistral.ai/inference/pricing", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Ministral 3 8B", "api_id": "ministral-3-8b", "tier": "small", "status": "ga", "release_date": "2025-12-02", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.15, "cached_input_price": 0.015, "output_price": 0.15, "price_notes": "Standardtarif. Batch kostet 50 % weniger; regionale Inferenz kostet für unterstützte Modelle 10 % mehr. Cache-Hits kosten 0,015 USD pro 1 Mio. Input-Tokens.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "batch", "prompt_caching" ], "open_weights": true, "license": "Apache 2.0", "description": "Kompaktes und effizientes Edge-Modell mit Text- und Vision-Fähigkeiten.", "source_urls": [ "https://docs.mistral.ai/models/ministral-3-8b-25-12", "https://docs.mistral.ai/inference/pricing", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Ministral 3 3B", "api_id": "ministral-3-3b", "tier": "small", "status": "ga", "release_date": "2025-12-02", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.1, "cached_input_price": 0.01, "output_price": 0.1, "price_notes": "Standardtarif. Batch kostet 50 % weniger; regionale Inferenz kostet für unterstützte Modelle 10 % mehr. Cache-Hits kosten 0,01 USD pro 1 Mio. Input-Tokens.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "batch", "prompt_caching" ], "open_weights": true, "license": "Apache 2.0", "description": "Kleinstes und besonders effizientes Modell der Ministral-3-Familie für lokale Text- und Vision-Aufgaben.", "source_urls": [ "https://docs.mistral.ai/models/ministral-3-3b-25-12", "https://docs.mistral.ai/inference/pricing", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Voxtral Small", "api_id": "voxtral-small", "tier": "specialized", "status": "ga", "release_date": "2025-07-15", "context_window": 32000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.1, "cached_input_price": null, "output_price": 0.4, "price_notes": "Der Text-Input kostet standardmäßig 0,10 USD pro 1 Mio. Tokens; Audio-Input wird getrennt mit 0,004 USD pro Minute abgerechnet. Batch kostet 50 % weniger; ein konkreter Cache-Tarif für dieses Modell ist nicht ausgewiesen.", "modalities_input": [ "text", "audio" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "batch" ], "open_weights": true, "license": "Apache 2.0", "description": "Multimodales Instruct-Modell mit Audioeingabe für Sprachverständnis und textbasierte Antworten.", "source_urls": [ "https://docs.mistral.ai/models/voxtral-small-25-07", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Z.ai GLM 5.2", "api_id": "zai-glm-5-2", "tier": "coding", "status": "preview", "release_date": "2026-08-06", "context_window": 1000000, "max_output_tokens": 128000, "knowledge_cutoff": null, "input_price": 1.4, "cached_input_price": 0.14, "output_price": 4.4, "price_notes": "Standardtarif für das von Mistral gehostete Drittanbieter-Modell. Batch kostet 50 % weniger; regionale Inferenz kostet für unterstützte Modelle 10 % mehr. Cache-Hits kosten 0,14 USD pro 1 Mio. Input-Tokens.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "batch", "prompt_caching" ], "open_weights": true, "license": null, "description": "Von Z.ai stammendes, unverändert von Mistral gehostetes Open-Source-Textmodell für Coding und agentische Long-Context-Workflows.", "source_urls": [ "https://docs.mistral.ai/models/zai-glm-5-2", "https://docs.mistral.ai/inference/pricing", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Leanstral 1.5", "api_id": "labs-leanstral-1-5", "tier": "coding", "status": "preview", "release_date": "2026-06-30", "context_window": 256000, "max_output_tokens": 128000, "knowledge_cutoff": null, "input_price": 0, "cached_input_price": null, "output_price": 0, "price_notes": "Der öffentliche Preview-Endpunkt ist derzeit kostenlos. Der Anbieter kündigt für dieses Modell die Einstellung zum 2026-09-30 an; es ist daher nicht für Produktionsbetrieb vorgesehen.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output" ], "open_weights": true, "license": "Apache 2.0", "description": "Preview-Modell für Lean-4-Proof-Engineering, automatisiertes Theorem Proving und Autoformalization.", "source_urls": [ "https://docs.mistral.ai/models/leanstral-1-5", "https://docs.mistral.ai/resources/changelogs", "https://mistral.ai/pricing/api/" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Mistral Medium 3.1", "api_id": "mistral-medium-3-1", "tier": "flagship", "status": "legacy", "release_date": "2025-08-12", "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Abgekündigt am 2026-05-22 und durch Mistral Medium 3.5 ersetzt. Die dokumentierte Frist für GA-Modelle beträgt sechs Monate; am Stichtag 2026-09-04 ist das Modell daher noch erreichbar, ein aktueller Preis ist jedoch nicht ausgewiesen.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "web_search", "code_execution", "batch" ], "open_weights": false, "license": null, "description": "Abgekündigtes multimodales Frontier-Modell der vorherigen Mistral-Medium-Generation.", "source_urls": [ "https://docs.mistral.ai/models/mistral-medium-3-1-25-08", "https://docs.mistral.ai/inference/model-lifecycle" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Magistral Medium 1.2", "api_id": "magistral-medium-1-2", "tier": "reasoning", "status": "legacy", "release_date": "2025-09-18", "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Abgekündigt am 2026-05-22 und für neue Integrationen durch Mistral Medium 3.5 ersetzt. Die dokumentierte Frist für GA-Modelle beträgt sechs Monate; am Stichtag 2026-09-04 ist das Modell daher noch erreichbar, ein aktueller Preis ist jedoch nicht ausgewiesen.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "batch" ], "open_weights": false, "license": null, "description": "Abgekündigtes Frontier-Multimodalmodell der Magistral-Reihe für Reasoning.", "source_urls": [ "https://docs.mistral.ai/models/magistral-medium-1-2-25-09", "https://docs.mistral.ai/inference/model-lifecycle" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Magistral Small 1.2", "api_id": "magistral-small-1-2", "tier": "reasoning", "status": "legacy", "release_date": "2025-09-18", "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Abgekündigt am 2026-04-30 und für neue Integrationen durch Mistral Small 4 ersetzt. Die dokumentierte Frist für GA-Modelle beträgt sechs Monate; am Stichtag 2026-09-04 ist das Modell daher noch erreichbar, ein aktueller Preis ist jedoch nicht ausgewiesen.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "batch" ], "open_weights": true, "license": "Apache 2.0", "description": "Abgekündigtes kleines multimodales Reasoning-Modell der Magistral-Reihe.", "source_urls": [ "https://docs.mistral.ai/models/magistral-small-1-2-25-09", "https://docs.mistral.ai/inference/model-lifecycle" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Devstral 2", "api_id": "devstral-2", "tier": "coding", "status": "legacy", "release_date": "2025-12-09", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Abgekündigt am 2026-05-22 und für neue Integrationen durch Mistral Medium 3.5 ersetzt. Die dokumentierte Frist für GA-Modelle beträgt sechs Monate; am Stichtag 2026-09-04 ist das Modell daher noch erreichbar, ein aktueller Preis ist jedoch nicht ausgewiesen.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "batch" ], "open_weights": true, "license": "Modified MIT", "description": "Abgekündigtes Open-Weights-Modell für agentische Softwareentwicklung mit Werkzeugnutzung und Bearbeitung mehrerer Dateien.", "source_urls": [ "https://docs.mistral.ai/models/devstral-2-25-12", "https://docs.mistral.ai/inference/model-lifecycle" ], "verified_in_api": null }, { "provider_id": "mistral", "provider": "Mistral AI", "name": "Mistral Small 3.2", "api_id": "mistral-small-3-2", "tier": "small", "status": "legacy", "release_date": "2025-06-20", "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Abgekündigt am 2026-04-30 und durch Mistral Small 4 ersetzt. Die dokumentierte Frist für GA-Modelle beträgt sechs Monate; am Stichtag 2026-09-04 ist das Modell daher noch erreichbar, ein aktueller Preis ist jedoch nicht ausgewiesen.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "web_search", "code_execution", "batch" ], "open_weights": true, "license": "Apache 2.0", "description": "Abgekündigtes kleines multimodales Allzweckmodell der vorherigen Mistral-Small-Generation.", "source_urls": [ "https://docs.mistral.ai/models/mistral-small-3-2-25-06", "https://docs.mistral.ai/inference/model-lifecycle" ], "verified_in_api": null }, { "provider_id": "moonshot", "provider": "Moonshot AI (Kimi)", "name": "Kimi K3", "api_id": "kimi-k3", "tier": "flagship", "status": "ga", "release_date": "2026-07-16", "context_window": 1048576, "max_output_tokens": 1048576, "knowledge_cutoff": null, "input_price": 3, "cached_input_price": 0.3, "output_price": 15, "price_notes": "Standardtarif bei Cache-Miss; Cache-Hits kosten 0,30 USD pro 1 Mio. Input-Tokens. Einheitlicher Tarif bis zum 1M-Kontextfenster, ohne Long-Context-Staffel; Preise zuzüglich ggf. anfallender Steuern.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "prompt_caching", "streaming" ], "open_weights": true, "license": "Kimi K3 License", "description": "Flaggschiff für langlaufende Coding- und Wissensarbeit sowie tiefes Reasoning mit nativem Bild- und Videoverständnis.", "source_urls": [ "https://platform.kimi.ai/docs/models", "https://platform.kimi.ai/docs/guide/kimi-k3-quickstart", "https://platform.kimi.ai/docs/pricing/chat-k3", "https://platform.kimi.ai/", "https://www.moonshot.ai/", "https://github.com/MoonshotAI/Kimi-K3/blob/main/LICENSE" ], "verified_in_api": true }, { "provider_id": "moonshot", "provider": "Moonshot AI (Kimi)", "name": "Kimi K2.6", "api_id": "kimi-k2.6", "tier": "standard", "status": "ga", "release_date": "2026-04-20", "context_window": 262144, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.95, "cached_input_price": 0.16, "output_price": 4, "price_notes": "Standardtarif bei Cache-Miss; Cache-Hits kosten 0,16 USD pro 1 Mio. Input-Tokens. Im Batch-Modus beträgt der Tarif 60 % des Echtzeit-Tarifs; Preise zuzüglich ggf. anfallender Steuern.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "batch", "prompt_caching", "streaming" ], "open_weights": true, "license": "Modified MIT License", "description": "Allzweckmodell für Dialog, Coding, Agentenaufgaben und multimodales Verstehen, dessen Thinking-Modus pro Anfrage deaktiviert werden kann.", "source_urls": [ "https://platform.kimi.ai/docs/models", "https://platform.kimi.ai/docs/guide/kimi-k2-6-quickstart", "https://platform.kimi.ai/docs/pricing/chat-k26", "https://www.kimi.ai/resources/kimi-k2-6-pricing", "https://huggingface.co/moonshotai/Kimi-K2.6/blob/main/README.md", "https://github.com/MoonshotAI/kimi-help-center/blob/master/en-US/agent/overview.md" ], "verified_in_api": true }, { "provider_id": "moonshot", "provider": "Moonshot AI (Kimi)", "name": "Kimi K2.7 Code HighSpeed", "api_id": "kimi-k2.7-code-highspeed", "tier": "coding", "status": "ga", "release_date": null, "context_window": 262144, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.9, "cached_input_price": 0.38, "output_price": 8, "price_notes": "Eigenständige HighSpeed-API-ID mit etwa 180 Tokens/s Ausgabe, in kurzen Kontexten bis zu 260 Tokens/s. Die Preise sind der Echtzeit-Standardtarif bei Cache-Miss; Cache-Hits kosten 0,38 USD pro 1 Mio. Input-Tokens; zuzüglich ggf. anfallender Steuern.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "Modified MIT License", "description": "Beschleunigte API-Variante von Kimi K2.7 Code für latenzkritische Coding-Workflows; Modellverhalten und Thinking-Modus entsprechen K2.7 Code.", "source_urls": [ "https://platform.kimi.ai/docs/models", "https://platform.kimi.ai/docs/guide/kimi-k2-7-code-quickstart", "https://platform.kimi.ai/docs/pricing/chat-k27-code", "https://www.kimi.ai/resources/kimi-k2-7-code-pricing", "https://huggingface.co/moonshotai/Kimi-K2.7-Code" ], "verified_in_api": true }, { "provider_id": "moonshot", "provider": "Moonshot AI (Kimi)", "name": "Kimi K2.7 Code", "api_id": "kimi-k2.7-code", "tier": "coding", "status": "ga", "release_date": null, "context_window": 262144, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 0.95, "cached_input_price": 0.19, "output_price": 4, "price_notes": "Standardtarif bei Cache-Miss; Cache-Hits kosten 0,19 USD pro 1 Mio. Input-Tokens. Im Batch-Modus beträgt der Tarif 60 % des Echtzeit-Tarifs; Preise zuzüglich ggf. anfallender Steuern.", "modalities_input": [ "text", "image", "video" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "batch", "prompt_caching", "streaming" ], "open_weights": true, "license": "Modified MIT License", "description": "Dediziertes Coding- und Agentenmodell für langlaufende Software-Engineering-Aufgaben mit dauerhaft aktiviertem Thinking.", "source_urls": [ "https://platform.kimi.ai/docs/models", "https://platform.kimi.ai/docs/guide/kimi-k2-7-code-quickstart", "https://platform.kimi.ai/docs/pricing/chat-k27-code", "https://www.kimi.ai/resources/kimi-k2-7-code-pricing", "https://huggingface.co/moonshotai/Kimi-K2.7-Code" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.6 Sol", "api_id": "gpt-5.6-sol", "tier": "flagship", "status": "ga", "release_date": "2026-07-09", "context_window": 1050000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-02-16", "input_price": 4, "cached_input_price": 0.4, "output_price": 20, "price_notes": "Standardtarif einschließlich der bis mindestens 2026-11-21 angekündigten Aktionspreise. Ab mehr als 272.000 Input-Tokens: 2× Input und 1,5× Output; Cache-Writes kosten 1,25× Input. Batch und Flex kosten 50 % des Standardtarifs, Fast Mode 2×; Regional Processing ggf. +10 %.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Flaggschiff der GPT-5.6-Familie für komplexe professionelle Aufgaben, Coding und agentische Workflows.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.6-sol", "https://platform.openai.com/docs/pricing", "https://openai.com/index/gpt-5-6/" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.6 Terra", "api_id": "gpt-5.6-terra", "tier": "standard", "status": "ga", "release_date": "2026-07-09", "context_window": 1050000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-02-16", "input_price": 2, "cached_input_price": 0.2, "output_price": 12, "price_notes": "Standardtarif. Ab mehr als 272.000 Input-Tokens: 2× Input und 1,5× Output; Cache-Writes kosten 1,25× Input. Batch und Flex kosten 50 % des Standardtarifs, Fast Mode 2×; Regional Processing ggf. +10 %.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Ausgewogenes GPT-5.6-Modell für Workloads, bei denen Intelligenz und Kosten im Gleichgewicht stehen sollen.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.6-terra", "https://platform.openai.com/docs/pricing", "https://openai.com/index/gpt-5-6/" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.3-Codex", "api_id": "gpt-5.3-codex", "tier": "coding", "status": "ga", "release_date": "2026-02-05", "context_window": 400000, "max_output_tokens": 128000, "knowledge_cutoff": "2025-08-31", "input_price": 1.75, "cached_input_price": 0.175, "output_price": 14, "price_notes": "Standardtarif. Fast Mode kostet 3,50 USD Input, 0,35 USD Cached Input und 28 USD Output pro 1 Mio. Tokens. Weitere Tool-Kosten, etwa für Web Search oder Container, werden separat abgerechnet.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Spezialisiertes Modell für agentisches Coding in Codex oder vergleichbaren Entwicklungsumgebungen.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.3-codex", "https://platform.openai.com/docs/pricing", "https://openai.com/index/introducing-gpt-5-3-codex/" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.6 Luna", "api_id": "gpt-5.6-luna", "tier": "small", "status": "ga", "release_date": "2026-07-09", "context_window": 1050000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-02-16", "input_price": 0.2, "cached_input_price": 0.02, "output_price": 1.2, "price_notes": "Standardtarif. Ab mehr als 272.000 Input-Tokens: 2× Input und 1,5× Output; Cache-Writes kosten 1,25× Input. Batch und Flex kosten 50 % des Standardtarifs, Fast Mode 2×; Regional Processing ggf. +10 %.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Das kostenoptimierte GPT-5.6-Modell für schnelle Anwendungen mit hohem Volumen.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.6-luna", "https://platform.openai.com/docs/pricing", "https://openai.com/index/gpt-5-6/" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.6 Cyber", "api_id": "gpt-5.6-cyber", "tier": "specialized", "status": "ga", "release_date": "2026-08-10", "context_window": 400000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-02-16", "input_price": 12.5, "cached_input_price": 1.25, "output_price": 75, "price_notes": "Standardtarif; Zugang setzt separate Genehmigung und Provisionierung im Daybreak-Programm voraus. Laut Modellspezifikation greifen oberhalb von 272.000 Input-Tokens 2× Input und 1,5× Output; Cache-Writes kosten 1,25× Input. Regional Processing ggf. +10 %.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Spezialmodell für autorisierte Schwachstellenforschung, Exploit-Validierung und Security-Tests durch zugelassene Verteidiger.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.6-cyber", "https://platform.openai.com/docs/pricing", "https://openai.com/index/expanding-daybreak-as-the-cyber-defense-window-narrows/" ], "verified_in_api": false }, { "provider_id": "openai", "provider": "OpenAI", "name": "gpt-oss-120b", "api_id": null, "tier": "specialized", "status": "ga", "release_date": "2025-08-05", "context_window": 131072, "max_output_tokens": 131072, "knowledge_cutoff": "2024-06-01", "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Open-Weights-Modell ohne separat bepreiste, nutzbare OpenAI-Hosted-API; die Infrastrukturkosten der eigenen Bereitstellung fallen außerhalb der OpenAI-Tokenpreise an.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "fine_tuning", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Leistungsstärkstes Open-Weights-Modell von OpenAI für lokale oder spezialisierte Reasoning- und Agent-Workloads.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-oss-120b", "https://openai.com/index/introducing-gpt-oss/" ], "verified_in_api": false }, { "provider_id": "openai", "provider": "OpenAI", "name": "gpt-oss-20b", "api_id": null, "tier": "specialized", "status": "ga", "release_date": "2025-08-05", "context_window": 131072, "max_output_tokens": 131072, "knowledge_cutoff": "2024-06-01", "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Open-Weights-Modell ohne separat bepreiste, nutzbare OpenAI-Hosted-API; die Infrastrukturkosten der eigenen Bereitstellung fallen außerhalb der OpenAI-Tokenpreise an.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "fine_tuning", "streaming" ], "open_weights": true, "license": "Apache-2.0", "description": "Kompakteres Open-Weights-Modell für latenzarme lokale und spezialisierte Anwendungen.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-oss-20b", "https://openai.com/index/introducing-gpt-oss/" ], "verified_in_api": false }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-6 Astra", "api_id": "gpt-6-astra", "tier": "flagship", "status": "preview", "release_date": "2026-09-03", "context_window": 1050000, "max_output_tokens": 128000, "knowledge_cutoff": "2026-04-30", "input_price": 10, "cached_input_price": 1, "output_price": 50, "price_notes": "Standardtarif. Ab mehr als 272.000 Input-Tokens gilt für die gesamte Anfrage der doppelte Input-/Cache-Tarif und der 1,5-fache Output-Tarif. Batch und Flex kosten 50 % des Standardtarifs; Fast Mode kostet das Doppelte. Regional Processing kostet, soweit verfügbar, 10 % Aufschlag.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Das leistungsstärkste Modell für besonders anspruchsvolle End-to-End-Arbeit, komplexes Reasoning, Coding, Recherche und Computer Use.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-6-astra", "https://platform.openai.com/docs/pricing", "https://openai.com/index/gpt-6-astra/" ], "verified_in_api": false }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.5", "api_id": "gpt-5.5", "tier": "flagship", "status": "legacy", "release_date": "2026-04-23", "context_window": 1050000, "max_output_tokens": 128000, "knowledge_cutoff": "2025-12-01", "input_price": 5, "cached_input_price": 0.5, "output_price": 30, "price_notes": "Standardtarif. Ab mehr als 272.000 Input-Tokens: 2× Input und 1,5× Output für die gesamte Session. Batch und Flex kosten 50 % des damaligen Standardtarifs; Priority/Fast-Verarbeitung wurde mit einem Aufschlag angeboten. Regional Processing kostet +10 %.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Vorgänger-Flaggschiff für anspruchsvolle professionelle Arbeit, agentisches Coding, Recherche und Computer Use.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.5", "https://openai.com/index/introducing-gpt-5-5/" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.4", "api_id": "gpt-5.4", "tier": "flagship", "status": "legacy", "release_date": "2026-03-05", "context_window": 1050000, "max_output_tokens": 128000, "knowledge_cutoff": "2025-08-31", "input_price": 2.5, "cached_input_price": 0.25, "output_price": 15, "price_notes": "Standardtarif. Ab mehr als 272.000 Input-Tokens: 2× Input und 1,5× Output für die gesamte Session. Regional Processing kostet +10 %; Batch-/Flex- und Fast-Tarife können abweichen.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Weiterhin verfügbares Vorgänger-Flaggschiff für professionelle Arbeit, Coding und Tool-gestützte agentische Aufgaben.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.4", "https://openai.com/index/introducing-gpt-5-4/" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.5 Pro", "api_id": "gpt-5.5-pro", "tier": "reasoning", "status": "legacy", "release_date": "2026-04-23", "context_window": 1050000, "max_output_tokens": 128000, "knowledge_cutoff": "2025-12-01", "input_price": 30, "cached_input_price": null, "output_price": 180, "price_notes": "Standardtarif; kein Rabatt für gecachten Input. Das Modell ist über die Responses API einschließlich Batch verfügbar; Regional Processing kostet +10 %.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch" ], "open_weights": false, "license": null, "description": "Rechenintensivere Pro-Variante von GPT-5.5 für besonders schwierige Aufgaben mit höherer Antwortgenauigkeit.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.5-pro", "https://openai.com/index/introducing-gpt-5-5/" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.4 Pro", "api_id": "gpt-5.4-pro", "tier": "reasoning", "status": "legacy", "release_date": "2026-03-05", "context_window": 1050000, "max_output_tokens": 128000, "knowledge_cutoff": "2025-08-31", "input_price": 30, "cached_input_price": null, "output_price": 180, "price_notes": "Standardtarif; kein Rabatt für gecachten Input. Ab mehr als 272.000 Input-Tokens: 2× Input und 1,5× Output für die gesamte Session. Regional Processing kostet +10 %.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "web_search", "computer_use", "batch", "streaming" ], "open_weights": false, "license": null, "description": "Pro-Variante von GPT-5.4 mit mehr Inferenz-Rechenaufwand für schwierigste professionelle Aufgaben.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.4-pro", "https://openai.com/index/introducing-gpt-5-4/" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.4 Mini", "api_id": "gpt-5.4-mini", "tier": "small", "status": "legacy", "release_date": "2026-03-17", "context_window": 400000, "max_output_tokens": 128000, "knowledge_cutoff": "2025-08-31", "input_price": 0.75, "cached_input_price": 0.075, "output_price": 4.5, "price_notes": "Standardtarif; Regional Processing kostet +10 %. Batch-, Flex- oder Fast-Varianten sind nicht als separate Modelle erfasst.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "computer_use", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Kleineres GPT-5.4-Modell für effiziente High-Volume-Workloads, Coding, Computer Use und Subagents.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.4-mini", "https://openai.com/index/introducing-gpt-5-4-mini-and-nano/" ], "verified_in_api": true }, { "provider_id": "openai", "provider": "OpenAI", "name": "GPT-5.4 nano", "api_id": "gpt-5.4-nano", "tier": "small", "status": "legacy", "release_date": "2026-03-17", "context_window": 400000, "max_output_tokens": 128000, "knowledge_cutoff": "2025-08-31", "input_price": 0.2, "cached_input_price": 0.02, "output_price": 1.25, "price_notes": "Standardtarif; Regional Processing kostet +10 %. Batch-, Flex- oder Fast-Varianten sind nicht als separate Modelle erfasst.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "web_search", "code_execution", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Preisgünstigstes GPT-5.4-Modell für Klassifikation, Extraktion, Ranking und einfache Subagent-Aufgaben.", "source_urls": [ "https://developers.openai.com/api/docs/models/gpt-5.4-nano", "https://openai.com/index/introducing-gpt-5-4-mini-and-nano/" ], "verified_in_api": true }, { "provider_id": "perplexity", "provider": "Perplexity (Sonar)", "name": "Sonar Pro", "api_id": "sonar-pro", "tier": "flagship", "status": "legacy", "release_date": "2025-01", "context_window": 200000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 3, "cached_input_price": null, "output_price": 15, "price_notes": "Zusätzlich zur Tokenabrechnung: Fast-Search-Request-Gebühr je 1.000 Anfragen von 6 USD (Low, Standard), 10 USD (Medium) oder 14 USD (High). Optionales Pro Search kostet 14 USD, 18 USD bzw. 22 USD je 1.000 Anfragen (Low/Medium/High); bei auto erfolgt die Abrechnung nach der automatischen Klassifikation. Bild-Inputs werden als Input-Tokens abgerechnet.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text", "image" ], "capabilities": [ "tool_use", "structured_output", "web_search", "streaming" ], "open_weights": false, "license": null, "description": "Premium-Suchmodell für komplexe Abfragen, vertiefte Informationssynthese und Pro Search mit mehrstufiger Werkzeugnutzung.", "source_urls": [ "https://docs.perplexity.ai/docs/sonar/models/sonar-pro", "https://docs.perplexity.ai/docs/getting-started/pricing", "https://docs.perplexity.ai/docs/sonar/pro-search/quickstart", "https://docs.perplexity.ai/docs/sonar/media", "https://docs.perplexity.ai/docs/resources/changelog" ], "verified_in_api": null }, { "provider_id": "perplexity", "provider": "Perplexity (Sonar)", "name": "Sonar Reasoning Pro", "api_id": "sonar-reasoning-pro", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 2, "cached_input_price": null, "output_price": 8, "price_notes": "Zusätzlich zur Tokenabrechnung: Request-Gebühr je 1.000 Anfragen von 6 USD (Low, Standard), 10 USD (Medium) oder 14 USD (High). Das Modell gibt Reasoning im -Abschnitt aus; Bild-Input in Verbindung mit Structured Outputs wird nicht unterstützt. Bild-Inputs werden ansonsten als Input-Tokens abgerechnet.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text", "image" ], "capabilities": [ "reasoning", "structured_output", "web_search", "streaming" ], "open_weights": false, "license": null, "description": "Reasoning-Modell für anspruchsvolle mehrstufige Analysen und Problemlösung mit Chain-of-Thought sowie Websuche.", "source_urls": [ "https://docs.perplexity.ai/docs/sonar/models/sonar-reasoning-pro", "https://docs.perplexity.ai/docs/getting-started/pricing", "https://docs.perplexity.ai/docs/sonar/media", "https://docs.perplexity.ai/docs/resources/changelog" ], "verified_in_api": null }, { "provider_id": "perplexity", "provider": "Perplexity (Sonar)", "name": "Sonar Deep Research", "api_id": "sonar-deep-research", "tier": "reasoning", "status": "legacy", "release_date": null, "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 2, "cached_input_price": null, "output_price": 8, "price_notes": "Neben 2 USD Input und 8 USD Output pro 1 Mio. Tokens werden Citation Tokens mit 2 USD pro 1 Mio. Tokens, Reasoning Tokens mit 3 USD pro 1 Mio. Tokens und Suchanfragen mit 5 USD pro 1.000 Suchanfragen berechnet. Der reasoning_effort kann low, medium oder high sein und beeinflusst insbesondere die verbrauchten Reasoning Tokens; asynchrone Ausführung ist verfügbar. Bild-Inputs werden als Input-Tokens abgerechnet.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text", "image" ], "capabilities": [ "reasoning", "structured_output", "web_search", "streaming" ], "open_weights": false, "license": null, "description": "Deep-Research-Modell für umfangreiche Webrecherchen über viele Quellen und detaillierte, fachlich fundierte Berichte.", "source_urls": [ "https://docs.perplexity.ai/docs/sonar/models/sonar-deep-research", "https://docs.perplexity.ai/docs/getting-started/pricing", "https://docs.perplexity.ai/docs/sonar/media", "https://docs.perplexity.ai/docs/resources/changelog" ], "verified_in_api": null }, { "provider_id": "perplexity", "provider": "Perplexity (Sonar)", "name": "Sonar", "api_id": "sonar", "tier": "small", "status": "legacy", "release_date": "2025-01", "context_window": 128000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1, "cached_input_price": null, "output_price": 1, "price_notes": "Zusätzlich zur Tokenabrechnung: Request-Gebühr je 1.000 Anfragen von 5 USD (Low, Standard), 8 USD (Medium) oder 12 USD (High) abhängig von der Search-Context-Größe. Bild-Inputs werden als Input-Tokens abgerechnet.", "modalities_input": [ "text", "image", "pdf" ], "modalities_output": [ "text", "image" ], "capabilities": [ "structured_output", "web_search", "streaming" ], "open_weights": false, "license": null, "description": "Kleines, schnelles und kostengünstiges Suchmodell für web-gestützte Antworten sowie einfache Frage-Antwort-Aufgaben.", "source_urls": [ "https://docs.perplexity.ai/docs/sonar/models/sonar", "https://docs.perplexity.ai/docs/getting-started/pricing", "https://docs.perplexity.ai/docs/sonar/media", "https://docs.perplexity.ai/docs/resources/changelog" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-5.3", "api_id": "glm-5.3", "tier": "flagship", "status": "ga", "release_date": "2026-08-18", "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 1.4, "cached_input_price": 0.26, "output_price": 4.4, "price_notes": "Standardtarif; Context-Cache-Speicherung laut Preisseite zeitlich begrenzt kostenlos.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "GLM-5.3 License", "description": "Das aktuelle Flaggschiff für anspruchsvolles Coding, lange agentische Aufgaben und steuerbare Reasoning-Tiefe.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/llm/glm-5.3", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-5.3", "https://github.com/zai-org/GLM-5" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-5.2", "api_id": "glm-5.2", "tier": "flagship", "status": "ga", "release_date": "2026-06-16", "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 1.4, "cached_input_price": 0.26, "output_price": 4.4, "price_notes": "Standardtarif; Context-Cache-Speicherung laut Preisseite zeitlich begrenzt kostenlos.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT", "description": "Ein Flaggschiff für langlaufende Coding-Agenten mit 1-Mio.-Token-Kontext und wählbarem Reasoning-Aufwand.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/llm/glm-5.2", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-5.2", "https://github.com/zai-org/GLM-5" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-5.1", "api_id": "glm-5.1", "tier": "flagship", "status": "ga", "release_date": "2026-04-07", "context_window": 200000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 1.4, "cached_input_price": 0.26, "output_price": 4.4, "price_notes": "Standardtarif; Context-Cache-Speicherung laut Preisseite zeitlich begrenzt kostenlos.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT", "description": "Ein weiterhin verfügbares GLM-5-Modell für langfristige agentische Engineering- und Coding-Aufgaben.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/overview/overview", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-5.1", "https://github.com/zai-org/GLM-5" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-5", "api_id": "glm-5", "tier": "flagship", "status": "ga", "release_date": "2026-02-12", "context_window": 200000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 1, "cached_input_price": 0.2, "output_price": 3.2, "price_notes": "Standardtarif; Context-Cache-Speicherung laut Preisseite zeitlich begrenzt kostenlos.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT", "description": "Ein Open-Weights-Modell für komplexes System-Engineering, tiefes Debugging und langfristige Agentenaufgaben.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/overview/overview", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-5", "https://github.com/zai-org/GLM-5" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-4.7", "api_id": "glm-4.7", "tier": "coding", "status": "ga", "release_date": "2025-12-22", "context_window": 200000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 0.6, "cached_input_price": 0.11, "output_price": 2.2, "price_notes": "Standardtarif; Context-Cache-Speicherung laut Preisseite zeitlich begrenzt kostenlos.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT", "description": "Ein Open-Weights-Coding- und Agentenmodell mit verbessertem mehrstufigem Reasoning und Tool-Use.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/llm/glm-4.7", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-4.7", "https://github.com/zai-org/GLM-4.5" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-4.7-FlashX", "api_id": "glm-4.7-flashx", "tier": "small", "status": "ga", "release_date": null, "context_window": 200000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 0.07, "cached_input_price": 0.01, "output_price": 0.4, "price_notes": "Standardtarif; Context-Cache-Speicherung laut Preisseite zeitlich begrenzt kostenlos.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Die schnelle und günstige GLM-4.7-Variante für allgemeine Agentic-Coding-Aufgaben mit hohem Durchsatz.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/llm/glm-4.7", "https://docs.z.ai/api-reference/llm/chat-completion" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-4.7-Flash", "api_id": "glm-4.7-flash", "tier": "small", "status": "ga", "release_date": "2026-01-19", "context_window": 200000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 0, "cached_input_price": 0, "output_price": 0, "price_notes": "Kostenloses Modell laut Preisseite.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT", "description": "Ein kostenloses, kompaktes Open-Weights-Modell für latenzsensible Coding-, Reasoning- und Textaufgaben.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/llm/glm-4.7", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-4.7-Flash", "https://github.com/zai-org/GLM-4.5" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-5.3-Flash", "api_id": "glm-5.3-flash", "tier": "specialized", "status": "ga", "release_date": "2026-08-26", "context_window": 1000000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 0.075, "cached_input_price": 0.015, "output_price": 0.25, "price_notes": "Aktueller, bis 2026-09-09 um 24:00 Uhr UTC+8 befristet um 50 % reduzierter Tarif; reguläre Listenpreise sind 0,15 USD Input, 0,03 USD Cached Input und 0,50 USD Output pro 1 Mio. Tokens. Context-Cache-Speicherung ist zeitlich begrenzt kostenlos.", "modalities_input": [ "text", "image", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT", "description": "Ein natives multimodales und besonders kosteneffizientes Modell für visuelles Coding, GUI-Aufgaben und lange Kontexte.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/vlm/glm-5.3-flash", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-5.3-Flash", "https://github.com/zai-org/GLM-5" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-4.6", "api_id": "glm-4.6", "tier": "coding", "status": "legacy", "release_date": "2025-09-30", "context_window": 200000, "max_output_tokens": 131072, "knowledge_cutoff": null, "input_price": 0.6, "cached_input_price": 0.11, "output_price": 2.2, "price_notes": "Standardtarif; Context-Cache-Speicherung laut Preisseite zeitlich begrenzt kostenlos.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT", "description": "Die weiterhin buchbare, letzte Vorgängergeneration für Coding, Reasoning und Tool-gestützte Agenten.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/overview/overview", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-4.6", "https://github.com/zai-org/GLM-4.5" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-4.6V", "api_id": "glm-4.6v", "tier": "specialized", "status": "legacy", "release_date": "2025-12-08", "context_window": 128000, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.3, "cached_input_price": 0.05, "output_price": 0.9, "price_notes": "Standardtarif; Context-Cache-Speicherung laut Preisseite zeitlich begrenzt kostenlos.", "modalities_input": [ "text", "image", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT", "description": "Ein Open-Weights-VLM für Bild-, Video-, Dokument- und GUI-Verständnis mit nativem multimodalem Function Calling.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/vlm/glm-4.6v", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-4.6V", "https://github.com/zai-org/GLM-V" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-4.6V-FlashX", "api_id": "glm-4.6v-flashx", "tier": "specialized", "status": "legacy", "release_date": "2025-12-08", "context_window": 128000, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0.04, "cached_input_price": 0.004, "output_price": 0.4, "price_notes": "Standardtarif; Context-Cache-Speicherung laut Preisseite zeitlich begrenzt kostenlos.", "modalities_input": [ "text", "image", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Die schnelle, preisgünstige multimodale GLM-4.6V-Endpunktvariante für visuelle Analyse und Tool-Aufgaben.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/vlm/glm-4.6v", "https://docs.z.ai/api-reference/llm/chat-completion" ], "verified_in_api": null }, { "provider_id": "zai", "provider": "Z.ai / Zhipu (GLM)", "name": "GLM-4.6V-Flash", "api_id": "glm-4.6v-flash", "tier": "specialized", "status": "legacy", "release_date": "2025-12-08", "context_window": 128000, "max_output_tokens": 32768, "knowledge_cutoff": null, "input_price": 0, "cached_input_price": 0, "output_price": 0, "price_notes": "Kostenloses Modell laut Preisseite.", "modalities_input": [ "text", "image", "video", "pdf" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "prompt_caching", "streaming" ], "open_weights": true, "license": "MIT", "description": "Ein kostenloses, kompaktes Open-Weights-VLM für multimodales Verständnis und visuell gestützte Tool-Use-Szenarien.", "source_urls": [ "https://docs.z.ai/guides/overview/pricing", "https://docs.z.ai/guides/vlm/glm-4.6v", "https://docs.z.ai/api-reference/llm/chat-completion", "https://huggingface.co/zai-org/GLM-4.6V-Flash", "https://github.com/zai-org/GLM-V" ], "verified_in_api": null }, { "provider_id": "xai", "provider": "xAI (Grok)", "name": "Grok 4.6", "api_id": "grok-4.6", "tier": "flagship", "status": "ga", "release_date": "2026-08-12", "context_window": 500000, "max_output_tokens": null, "knowledge_cutoff": "2026-02-01", "input_price": 2.0, "cached_input_price": 0.5, "output_price": 6.0, "price_notes": "Ab 200.000 Prompt-Tokens gelten für alle Tokens des Requests 4,00 $ Input, 1,00 $ Cached Input und 12,00 $ Output pro 1 Mio. Tokens. Priority Processing kostet das Doppelte der Standardpreise; Batch wird nicht unterstützt.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Aktuelles Flaggschiff für Coding, agentische Aufgaben und Wissensarbeit mit konfigurierbarer Reasoning-Tiefe.", "source_urls": [ "https://docs.x.ai/developers/models/grok-4.6", "https://docs.x.ai/developers/pricing", "https://docs.x.ai/developers/release-notes", "https://x.ai/news/grok-4-6" ], "verified_in_api": null }, { "provider_id": "xai", "provider": "xAI (Grok)", "name": "Grok 4.3", "api_id": "grok-4.3", "tier": "standard", "status": "ga", "release_date": null, "context_window": 1000000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.25, "cached_input_price": 0.2, "output_price": 2.5, "price_notes": "Ab 200.000 Prompt-Tokens gelten 2,50 $ Input, 0,40 $ Cached Input und 5,00 $ Output pro 1 Mio. Tokens. Priority Processing kostet das Doppelte; die Batch API wird unterstützt und gewährt 20 % Rabatt.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Schnelles Standardmodell mit starkem Tool Calling, Instruction Following und abschalt- beziehungsweise konfigurierbarem Reasoning.", "source_urls": [ "https://docs.x.ai/developers/models/grok-4.3", "https://docs.x.ai/developers/pricing", "https://docs.x.ai/developers/migration/may-15-retirement" ], "verified_in_api": null }, { "provider_id": "xai", "provider": "xAI (Grok)", "name": "Grok Build 0.1", "api_id": "grok-build-0.1", "tier": "coding", "status": "preview", "release_date": "2026-05-19", "context_window": 256000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.0, "cached_input_price": 0.2, "output_price": 2.0, "price_notes": "Der Coding-Modellzugang befindet sich laut Release Notes weiterhin im Early Access. Ab 200.000 Prompt-Tokens gelten 2,00 $ Input, 0,40 $ Cached Input und 4,00 $ Output pro 1 Mio. Tokens; Priority Processing kostet das Doppelte, Batch wird nicht unterstützt.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Early-Access-Codingmodell für agentische Softwareentwicklung, Engineering- und Workflow-Aufgaben.", "source_urls": [ "https://docs.x.ai/developers/models/grok-build-0.1", "https://docs.x.ai/developers/pricing", "https://docs.x.ai/developers/release-notes" ], "verified_in_api": null }, { "provider_id": "xai", "provider": "xAI (Grok)", "name": "Grok 4.20 Multi-Agent Beta", "api_id": "grok-4.20-multi-agent", "tier": "specialized", "status": "preview", "release_date": "2026-03-10", "context_window": 1000000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.25, "cached_input_price": 0.2, "output_price": 2.5, "price_notes": "Die undatierte API-ID ist ein Alias des dokumentierten Snapshots `grok-4.20-multi-agent-0309`. Ab 200.000 Prompt-Tokens gelten 2,50 $ Input, 0,40 $ Cached Input und 5,00 $ Output pro 1 Mio. Tokens; Priority Processing kostet das Doppelte und Batch bietet 20 % Rabatt.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Beta-Spezialmodell, bei dem mehrere Agenten parallel zusammenarbeiten und ein Leader-Agent das Ergebnis synthetisiert.", "source_urls": [ "https://docs.x.ai/developers/models/grok-4.20-multi-agent-0309", "https://docs.x.ai/developers/pricing", "https://docs.x.ai/developers/release-notes", "https://docs.x.ai/developers/model-capabilities/text/multi-agent" ], "verified_in_api": null }, { "provider_id": "xai", "provider": "xAI (Grok)", "name": "Grok 4.5", "api_id": "grok-4.5", "tier": "flagship", "status": "legacy", "release_date": "2026-07-08", "context_window": 500000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 2.0, "cached_input_price": 0.3, "output_price": 6.0, "price_notes": "Weiterhin direkt buchbarer Vorgänger des Flaggschiffs. Ab 200.000 Prompt-Tokens gelten 4,00 $ Input, 0,60 $ Cached Input und 12,00 $ Output pro 1 Mio. Tokens; Priority Processing kostet das Doppelte, Batch wird nicht unterstützt.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Vorgänger-Flaggschiff für Coding, agentische Aufgaben und Wissensarbeit mit konfigurierbarem Reasoning.", "source_urls": [ "https://docs.x.ai/developers/models/grok-4.5", "https://docs.x.ai/developers/pricing", "https://docs.x.ai/developers/release-notes" ], "verified_in_api": null }, { "provider_id": "xai", "provider": "xAI (Grok)", "name": "Grok-1", "api_id": null, "tier": "flagship", "status": "legacy", "release_date": "2024-03-17", "context_window": 8192, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": null, "cached_input_price": null, "output_price": null, "price_notes": "Keine eigene gehostete xAI-API und daher keine tokenbasierten API-Preise; die Gewichte und Architektur wurden zur lokalen beziehungsweise selbst betriebenen Nutzung veröffentlicht.", "modalities_input": [ "text" ], "modalities_output": [ "text" ], "capabilities": [], "open_weights": true, "license": "Apache-2.0", "description": "Offen veröffentlichter 314B-MoE-Basiskontrollpunkt, der nicht für Dialoganwendungen nachtrainiert wurde.", "source_urls": [ "https://x.ai/news/grok-os", "https://github.com/xai-org/grok-1" ], "verified_in_api": null }, { "provider_id": "xai", "provider": "xAI (Grok)", "name": "Grok 4.20 (Reasoning)", "api_id": "grok-4.20-reasoning", "tier": "reasoning", "status": "legacy", "release_date": "2026-03-10", "context_window": 1000000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.25, "cached_input_price": 0.2, "output_price": 2.5, "price_notes": "Die undatierte API-ID ist ein Alias des aktuell dokumentierten Snapshots `grok-4.20-0309-reasoning`. Ab 200.000 Prompt-Tokens gelten 2,50 $ Input, 0,40 $ Cached Input und 5,00 $ Output pro 1 Mio. Tokens; Priority Processing kostet das Doppelte und Batch bietet 20 % Rabatt.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "reasoning", "tool_use", "structured_output", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Weiter buchbare Reasoning-Variante der Grok-4.20-Generation für präzise agentische und Tool-Calling-Aufgaben.", "source_urls": [ "https://docs.x.ai/developers/models/grok-4.20", "https://docs.x.ai/developers/pricing", "https://docs.x.ai/developers/release-notes" ], "verified_in_api": null }, { "provider_id": "xai", "provider": "xAI (Grok)", "name": "Grok 4.20 (Non-Reasoning)", "api_id": "grok-4.20-non-reasoning", "tier": "standard", "status": "legacy", "release_date": "2026-03-10", "context_window": 1000000, "max_output_tokens": null, "knowledge_cutoff": null, "input_price": 1.25, "cached_input_price": 0.2, "output_price": 2.5, "price_notes": "Die undatierte API-ID ist ein Alias des aktuell dokumentierten Snapshots `grok-4.20-0309-non-reasoning`. Ab 200.000 Prompt-Tokens gelten 2,50 $ Input, 0,40 $ Cached Input und 5,00 $ Output pro 1 Mio. Tokens; Priority Processing kostet das Doppelte und Batch bietet 20 % Rabatt.", "modalities_input": [ "text", "image" ], "modalities_output": [ "text" ], "capabilities": [ "tool_use", "structured_output", "batch", "prompt_caching", "streaming" ], "open_weights": false, "license": null, "description": "Weiter buchbare Grok-4.20-Variante für Antworten ohne aktiviertes Reasoning.", "source_urls": [ "https://docs.x.ai/developers/models/grok-4.20-beta-0309-non-reasoning", "https://docs.x.ai/developers/pricing", "https://docs.x.ai/developers/release-notes" ], "verified_in_api": null } ], "live_api_lists": { "openai": [ "babbage-002", "chat-latest", "chatgpt-image-latest", "davinci-002", "gpt-3.5-turbo", "gpt-3.5-turbo-0125", "gpt-3.5-turbo-1106", "gpt-3.5-turbo-16k", "gpt-3.5-turbo-instruct", "gpt-3.5-turbo-instruct-0914", "gpt-4", "gpt-4-0613", "gpt-4-turbo", "gpt-4-turbo-2024-04-09", "gpt-4.1", "gpt-4.1-2025-04-14", "gpt-4.1-mini", "gpt-4.1-mini-2025-04-14", "gpt-4.1-nano", "gpt-4.1-nano-2025-04-14", "gpt-4o", "gpt-4o-2024-05-13", "gpt-4o-2024-08-06", "gpt-4o-2024-11-20", "gpt-4o-mini", "gpt-4o-mini-2024-07-18", "gpt-4o-mini-search-preview", "gpt-4o-mini-search-preview-2025-03-11", "gpt-4o-mini-transcribe", "gpt-4o-mini-transcribe-2025-03-20", "gpt-4o-mini-transcribe-2025-12-15", "gpt-4o-mini-tts", "gpt-4o-mini-tts-2025-03-20", "gpt-4o-mini-tts-2025-12-15", "gpt-4o-search-preview", "gpt-4o-search-preview-2025-03-11", "gpt-4o-transcribe", "gpt-4o-transcribe-diarize", "gpt-5", "gpt-5-2025-08-07", "gpt-5-chat-latest", "gpt-5-codex", "gpt-5-mini", "gpt-5-mini-2025-08-07", "gpt-5-nano", "gpt-5-nano-2025-08-07", "gpt-5-pro", "gpt-5-pro-2025-10-06", "gpt-5-search-api", "gpt-5-search-api-2025-10-14", "gpt-5.1", "gpt-5.1-2025-11-13", "gpt-5.1-chat-latest", "gpt-5.1-codex", "gpt-5.1-codex-max", "gpt-5.1-codex-mini", "gpt-5.2", "gpt-5.2-2025-12-11", "gpt-5.2-chat-latest", "gpt-5.2-codex", "gpt-5.2-pro", "gpt-5.2-pro-2025-12-11", "gpt-5.3-chat-latest", "gpt-5.3-codex", "gpt-5.4", "gpt-5.4-2026-03-05", "gpt-5.4-mini", "gpt-5.4-mini-2026-03-17", "gpt-5.4-nano", "gpt-5.4-nano-2026-03-17", "gpt-5.4-pro", "gpt-5.4-pro-2026-03-05", "gpt-5.5", "gpt-5.5-2026-04-23", "gpt-5.5-pro", "gpt-5.5-pro-2026-04-23", "gpt-5.6-luna", "gpt-5.6-sol", "gpt-5.6-terra", "gpt-audio", "gpt-audio-1.5", "gpt-audio-2025-08-28", "gpt-audio-mini", "gpt-audio-mini-2025-10-06", "gpt-audio-mini-2025-12-15", "gpt-image-1", "gpt-image-1-mini", "gpt-image-1.5", "gpt-image-2", "gpt-image-2-2026-04-21", "gpt-live-transcribe", "gpt-realtime", "gpt-realtime-1.5", "gpt-realtime-2", "gpt-realtime-2.1", "gpt-realtime-2.1-mini", "gpt-realtime-2025-08-28", "gpt-realtime-mini", "gpt-realtime-mini-2025-12-15", "gpt-realtime-translate", "gpt-realtime-whisper", "gpt-transcribe", "o1", "o1-2024-12-17", "o1-pro", "o1-pro-2025-03-19", "o3", "o3-2025-04-16", "o3-mini", "o3-mini-2025-01-31", "o4-mini", "o4-mini-2025-04-16", "omni-moderation-2024-09-26", "omni-moderation-latest", "sora-2", "sora-2-pro", "text-embedding-3-large", "text-embedding-3-small", "text-embedding-ada-002", "tts-1", "tts-1-1106", "tts-1-hd", "tts-1-hd-1106", "whisper-1" ], "anthropic": [ "claude-fable-5", "claude-fable-5-1", "claude-haiku-4-5-20251001", "claude-opus-4-5-20251101", "claude-opus-4-6", "claude-opus-4-7", "claude-opus-4-8", "claude-opus-5", "claude-sonnet-4-5-20250929", "claude-sonnet-4-6", "claude-sonnet-5" ], "gemini": [ "antigravity-preview-05-2026", "aqa", "deep-research-max-preview-04-2026", "deep-research-preview-04-2026", "deep-research-pro-preview-12-2025", "gemini-2.5-computer-use-preview-10-2025", "gemini-2.5-flash", "gemini-2.5-flash-image", "gemini-2.5-flash-lite", "gemini-2.5-flash-native-audio-latest", "gemini-2.5-flash-native-audio-preview-09-2025", "gemini-2.5-flash-native-audio-preview-12-2025", "gemini-2.5-flash-preview-tts", "gemini-2.5-pro", "gemini-2.5-pro-preview-tts", "gemini-3-flash-preview", "gemini-3-pro-image", "gemini-3-pro-image-preview", "gemini-3.1-flash-image", "gemini-3.1-flash-image-preview", "gemini-3.1-flash-lite", "gemini-3.1-flash-lite-image", "gemini-3.1-flash-lite-preview", "gemini-3.1-flash-live-preview", "gemini-3.1-flash-tts-preview", "gemini-3.1-pro-preview", "gemini-3.1-pro-preview-customtools", "gemini-3.5-flash", "gemini-3.5-flash-lite", "gemini-3.5-live-translate-preview", "gemini-3.5-transcribe", "gemini-3.5-transcribe-live", "gemini-3.6-flash", "gemini-3.7-flash", "gemini-3.8-flash", "gemini-embedding-001", "gemini-embedding-2", "gemini-embedding-2-preview", "gemini-flash-latest", "gemini-flash-lite-latest", "gemini-omni-1.1-flash", "gemini-omni-flash-preview", "gemini-pro-latest", "gemini-robotics-er-2-preview", "gemini-robotics-er-2-streaming-preview", "gemma-4-26b-a4b-it", "gemma-4-31b-it", "lyria-3-clip-preview", "lyria-3-pro-preview", "lyria-3.5", "nano-banana-pro-preview", "veo-3.1-fast-generate-preview", "veo-3.1-generate-preview", "veo-3.1-lite-generate-preview" ], "deepseek": [ "deepseek-v4-flash", "deepseek-v4-flash-vision-exp", "deepseek-v4-pro" ], "moonshot": [ "kimi-k2.6", "kimi-k2.7-code", "kimi-k2.7-code-highspeed", "kimi-k3" ] } }