Models

All models available through the API. Use the same model ID in your request to access any model.

Uso: "model": "qwen3.8-max"

Pass the model name in the "model" field of your request body.

Capabilities

ChatReasoningVisionToolsImage

Model	Provider	Capabilities	Speed	Context
`qwen3.8-max`	Qwen	ChatReasoningVision	Fast	128K
`qwen3-max`	Qwen	ChatReasoning	Medium	128K
`deepseek-v4-flash`	DeepSeek	ChatReasoning	Fast	128K
`deepseek-r1`	DeepSeek	ChatReasoning	Slow	128K
`gpt-4o`	OpenAI	ChatVisionTools	Medium	128K
`gpt-4o-mini`	OpenAI	ChatVisionTools	Fast	128K
`o3-mini`	OpenAI	ChatReasoning	Medium	128K
`gemini-2.5-flash`	Google	ChatVisionReasoning	Fast	1M
`gemini-2.5-pro`	Google	ChatVisionReasoning	Medium	1M
`llama-4-maverick`	Meta	ChatVision	Fast	128K
`llama-4-scout`	Meta	Chat	Fast	128K
`llama-3.1-nemotron-70b`	NVIDIA	ChatReasoning	Medium	128K
`command-a`	Cohere	ChatTools	Fast	128K
`kimi-k2`	Kimi	ChatReasoningVision	Medium	128K
`grok-3`	Grok	ChatReasoningVision	Medium	128K
`dall-e-3`	OpenAI	Image	Medium	N/A
`imagen-4`	Google	Image	Fast	N/A

The model list is updated regularly. Check /models for the most current list.

Models with "reasoning" capability can use enable_thinking to skip the reasoning step.

Image models (DALL-E, Imagen) use the /v1/image endpoint instead of /v1/chat/completions.