MODELS
by MiniMax
MiniMax M3 is a MiniMax frontier multimodal coding model with a 1M-token context window for agentic reasoning, tool use, coding, multimodal chat input, and long-context tasks.
MiniMax M3 is listed by MiniMax as a coding and agentic frontier model with 1M-context MSA and native multimodality. MiniMax API documentation describes MiniMax-M3 as a frontier multimodal coding model with a 1,000,000-token context window and documents Anthropic- and OpenAI-compatible API calls using the MiniMax-M3 model ID. MiniMax release notes say it was released for agentic reasoning, tool use, coding, multimodal chat input, and long-context tasks. The OpenRouter listing describes MiniMax-M3 as supporting text, image, and video inputs with text output and being suited for long-horizon agentic work and coding.
Use for coding workflows where MiniMax’s documented focus on frontier coding code understanding multi-turn dialogue and reasoning is relevant. Use for agentic reasoning tool-use or long-horizon agent workflows that benefit from a 1M-token context window. Use for multimodal chat/input scenarios involving text image or video inputs with text output as described in supplied model listings.
Last updated Jun 1, 2026
