MiniMax M2.5

MiniMax-M2.5 is an MoE model (230B total) extensively trained with reinforcement learning in hundreds of thousands of real-world environments, delivering SOTA results in coding, agentic tool use, search, and office work.

Highlights

SOTA Coding. M2.5 was trained on over 10 languages across more than 200,000 real-world environments. Going far beyond bug-fixing, M2.5 delivers reliable performance across the entire development lifecycle of complex systems: from 0-to-1 system design and environment setup, to 1-to-10 system development, to 10-to-90 feature iteration, and finally 90-to-100 comprehensive code review and system testing. It covers full-stack projects spanning multiple platforms including Web, Android, iOS, and Windows.

Search & Tool Use. Effective tool calling and search are prerequisites for a model's ability to autonomously handle more complex tasks. M2.5 achieved industry-leading performance on benchmarks such as BrowseComp and Wide Search. The model's generalization has also improved — M2.5 demonstrates more stable performance when facing unfamiliar scaffolding environments.

Agent Performance. M2.5 demonstrates much better decision-making when handling agentic tasks: it has learned to solve problems with more precise search rounds and better token efficiency. A significant improvement is M2.5's ability to think and plan like an architect — before writing any code, M2.5 actively decomposes and plans the features, structure, and UI design of the project from the perspective of an experienced software architect.

Office Work. M2.5 was trained to produce truly deliverable outputs in office scenarios through thorough collaboration with senior professionals in fields such as finance, law, and social sciences. M2.5 has achieved significant capability improvements in high-value workspace scenarios such as Word, PowerPoint, and Excel financial modeling.