Model catalog

Whisper Large v3 TurboOpenAI

Whisper Large v3 Turbo accepts audio and returns text. The canonical context window is 448 tokens. The canonical maximum output is 448 tokens.

Llama-3.2-1BMeta

Llama-3.2-1B accepts text and returns text. The canonical context window is 131,072 tokens. The canonical maximum output is 8,192 tokens.

Llama-3.2-3BMeta

Llama-3.2-3B accepts text and returns text. The canonical context window is 131,072 tokens. The canonical maximum output is 8,192 tokens.

Llama-3.2-11B-Vision-InstructMeta

Llama-3.2-11B-Vision-Instruct accepts text, images and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 4,096 tokens.

Llama 3.2 1B InstructMeta

Llama 3.2 1B Instruct accepts text and returns text. The canonical context window is 131,072 tokens. The canonical maximum output is 32,000 tokens.

Pixtral 12BMistral

Pixtral 12B accepts text, images and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 128,000 tokens.

Qwen2.5-VL 72B InstructAlibaba

Qwen2.5-VL 72B Instruct accepts text, images and returns text. The canonical context window is 131,072 tokens. The canonical maximum output is 8,192 tokens.

Command RCohere

Command R accepts text and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 4,000 tokens.

Command R+Cohere

Command R+ accepts text and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 4,000 tokens.

Llama 3.1 Euryale 70B v2.2Sao10k

Llama 3.1 Euryale 70B v2.2 accepts text and returns text. The canonical context window is 131,072 tokens. The canonical maximum output is 16,384 tokens.

Hermes 3 70B InstructNousresearch

Hermes 3 70B Instruct accepts text and returns text. The canonical context window is 131,072 tokens. The canonical maximum output is 16,384 tokens.

Hermes 3 405B InstructNousresearch

Hermes 3 405B Instruct accepts text and returns text. The canonical context window is 131,072 tokens. The canonical maximum output is 16,384 tokens.

Llama 3 8B LunarisSao10k

Llama 3 8B Lunaris accepts text and returns text. The canonical context window is 8,192 tokens. The canonical maximum output is 7,372 tokens.

GPT-4o (2024-08-06)OpenAI

GPT-4o (2024-08-06) accepts text, images and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 16,384 tokens.

Llama-3.1-8B-InstructMeta

Llama-3.1-8B-Instruct accepts text and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 4,096 tokens.

Llama-3.1-70B-InstructMeta

Llama-3.1-70B-Instruct accepts text and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 4,096 tokens.

Llama-Guard-3-8BMeta

Llama-Guard-3-8B accepts text and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 4,096 tokens.

GPT-4o miniOpenAI

GPT-4o mini accepts text, images, PDFs and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 16,384 tokens.

Gemma 2 27BGoogle

Gemma 2 27B accepts text and returns text. The canonical context window is 8,192 tokens. The canonical maximum output is 2,048 tokens.

Mistral NemoMistral

Mistral Nemo accepts text and returns text. The canonical context window is 128,000 tokens. The canonical maximum output is 128,000 tokens.

Codestral (latest)Mistral

Codestral (latest) accepts text and returns text. The canonical context window is 256,000 tokens. The canonical maximum output is 4,096 tokens.

Codestral-22B-v0.1Mistral

Codestral-22B-v0.1 accepts text and returns text. The canonical context window is 32,768 tokens. The canonical maximum output is 8,192 tokens.

Vision LargeVispark

Vision Large accepts text, images, audio, video, PDFs and returns text. The canonical context window is 1,000,000 tokens. The canonical maximum output is 65,536 tokens.

Vision MediumVispark

Vision Medium accepts text, images, audio, video, PDFs and returns text. The canonical context window is 1,000,000 tokens. The canonical maximum output is 65,536 tokens.