Access 9 text models and 3 media models through a single API.

Text Models

Model	ID	Provider	Vision	Specialty
Claude	`claude`	Anthropic	Yes	Thoughtful, analytical reasoning
GPT	`gpt`	OpenAI	Yes	Creative problem-solving
Grok	`grok`	xAI	Yes	Bold, unconventional perspectives
Gemini	`gemini`	Google	Yes	Comprehensive knowledge
DeepSeek	`deepseek`	DeepSeek	No	Technical depth, coding
Kimi	`kimi`	Moonshot	No	Multilingual capabilities
Mistral	`mistral`	Mistral AI	No	Efficient, practical insights
Llama	`llama`	Meta	No	Open-source AI
MiniMax	`minimax`	MiniMax	No	Strong multilingual, roleplay

Media Models (MiniMax)

Model	ID	Type	Description
Speech 2.6	`speech-2.6-hd`	TTS	Text-to-speech, 300+ voices, 40 languages
Music 2.5	`music-2.5`	Music	Music generation from text/lyrics
Hailuo 2.3	`MiniMax-Hailuo-2.3`	Video	AI video generation with camera control

Model Aliases

Each model supports multiple aliases for convenience.

Claude (Anthropic)

claude, anthropic, claude-sonnet-4-5, claude-sonnet-4-6, claude-opus-4-5, claude-opus-4-6

GPT (OpenAI) — includes reasoning models

gpt, chatgpt, openai, gpt-4, gpt-4o, gpt-5, gpt-5.2, o3, o3-mini, o4, o4-mini

Grok (xAI)

grok, x-ai, grok-3, grok-4, grok-4.1, grok-4-1-fast, grok-4-1-fast-reasoning

Gemini (Google) — includes reasoning mode

gemini, google, gemini-pro, gemini-2.5-pro, gemini-2.5-flash, gemini-3-flash, gemini-3-pro

DeepSeek — includes reasoning (R1)

deepseek, deepseek-chat, deepseek-v3, deepseek-v3.2, deepseek-r1, deepseek-reasoner

Kimi (Moonshot)

kimi, moonshot, kimi-k2, kimi-k2.5, moonshot-v1-128k

Mistral — includes code (Codestral)

mistral, mistral-large, mistral-large-latest, ministral-14b, codestral, codestral-latest

Llama (Meta)

llama, meta, llama-3, llama-3.3, llama-4, llama-4-maverick

Qwen (via OpenRouter → DeepSeek provider)

qwen3-coder, qwen3-coder-480b, qwen-coder

MiniMax

minimax, MiniMax-M2.7, minimax-m2.7, MiniMax-M2.7-highspeed, M2-her

Vision Support

Four models natively support image analysis. Non-vision models automatically receive an AI-generated description (produced once per request by GPT-5.2, then shared across all non-vision models).

Model	Native Vision	What Happens
GPT	Yes	Sees the image directly
Claude	Yes	Sees the image directly
Gemini	Yes	Sees the image directly
Grok	Yes	Sees the image directly
DeepSeek	No	Receives AI-generated description
Kimi	No	Receives AI-generated description
Mistral	No	Receives AI-generated description
Llama	No	Receives AI-generated description

Vision works in both the Gateway API (via OpenAI multimodal format) and the Chat API (via image_url parameter). The system handles capabilities automatically — you don't need to check which models support vision.

List Models (API)

GET /api/v1/models

Returns all available models in OpenAI-compatible format.

{
  "object": "list",
  "data": [
    {
      "id": "gpt-5.2",
      "object": "model",
      "created": 1706540000,
      "owned_by": "concurred/openai",
      "_concurred": {
        "name": "GPT-5.2",
        "aliases": ["chatgpt", "gpt-4", "gpt-5.2"]
      }
    }
  ]
}

Models

On this page