MiniMax-M1
A self-hosted reasoning LLM with a million-token context window and function calling, deployable through vLLM or Transformers.
MiniMax-M1 is a reasoning model for developers building AI agents or working on coding and mathematical problems with long inputs. You can deploy it on your own infrastructure.
- Long context: The model accepts up to a million tokens of context for tasks that need large inputs and extended reasoning.
- Coding and mathematics: Its training covers mathematical reasoning and software engineering tasks in sandbox environments; it also handles summarization, translation, question answering and creative writing.
- Function calling: It can request external functions and supply structured arguments, so developers can connect an agent to tools.
MiniMax-M1 builds on MiniMax-Text-01 and uses a mixture-of-experts architecture with lightning attention to reduce computation during long generations. Hugging Face hosts MiniMax-M1-40k and MiniMax-M1-80k, which have different thinking budgets.
Self-hosted deployment supports vLLM, the recommended production backend, or Transformers. The repository uses the Apache 2.0 license. MiniMax also offers a hosted chatbot with online search and an online API; these are separate from deployment on your own servers.
Logotyper
