MiniMax M3.1 Flash
M3.1 Flash builds on M3 with sparse attention throughout the model, quantization-aware training and an included DSpark draft head. It accepts text, images and video, then produces text.
The speed changes also affect local deployment
The card identifies three efficiency changes rather than a smaller parameter count: sparse attention across all layers, quantization-aware training, and speculative decoding. The DSpark head is part of the release. The usable context still depends on GPU memory and the serving setup.MiniMax M3.1 Flash model card
Check the full agent loop
Practical guidance. Compare completed tasks at a fixed effort setting. Include tool calls, image handling and time to a correct result. A fast token stream is not enough if the agent repeats a failed action or leaves work incomplete.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | API, Downloadable, Provider product |
| Model identifiers | MiniMax-M3.1-Flash |
| Context window | 1,000,000 tokens Maximum context stated for M3.1 Flash; deployment resources can limit usable length. MiniMax M3.1 Flash model card |
| Architecture | About 428 billion total parameters, with 23 billion active per token.MiniMax M3.1 Flash model card |
| Reasoning control | The API offers low, medium, high, xhigh and max effort. Thinking cannot be disabled; incompatible settings return an error.MiniMax text model invocation and thinking controls |
| Weight terms | The Community License requires commercial attribution and a one-time notice. Products or services above US$20 million yearly revenue require prior written authorization.MiniMax M3.1 Flash Community License |
Related models
- MiniMax M3
MiniMax · MiniMax M3
A native multimodal model with sparse attention that accepts text, images and video.
Launch date API · Downloadable · Provider product - MiniMax M2.7
MiniMax · MiniMax M2
A model for software agents and editable office work, including multi-agent teams and dynamic tool search.
Launch date Deprecated API; check access detailsAPI · Downloadable