MiniMax-M2.7
MiniMax-M2.7 is an open-weight language model from MiniMax, released on March 18, 2026.
- Developer
- MiniMax
- Released
- March 18, 2026
- Total parameters
- 229B
- Architecture
- MoE
- Input
- Text
- License
- MiniMax Non-Commercial License
- Commercial use
- No
- Hugging Face repository
- huggingface.co
- Checked on
MiniMax-M2.7 is a mixture-of-experts model. MiniMax does not state the parameter count on the model card; Hugging Face metadata counts about 229B parameters. The model takes text input. The weights are released under the MiniMax Non-Commercial License, which does not allow commercial use. Personal, non-profit and academic research use only. Any commercial use needs prior written authorization from MiniMax and the attribution “Built with MiniMax M2.7”.
- Total parameters, billions
- 229B
- License type
- Non-commercial
- License conditions
- Personal, non-profit and academic research use only. Any commercial use needs prior written authorization from MiniMax and the attribution "Built with MiniMax M2.7".
- Weights precision
- FP8 + BF16
- License text
- github.com
- Sources
Page URL Hugging Face model card https://huggingface.co/MiniMaxAI/MiniMax-M2.7/raw/main/README.md License text on GitHub https://raw.githubusercontent.com/MiniMax-AI/MiniMax-M2.7/main/LICENSE Hugging Face API https://huggingface.co/api/models/MiniMaxAI/MiniMax-M2.7?expand=safetensors minimax.io announcement https://www.minimax.io/news/minimax-m27-en
Questions
What license does MiniMax-M2.7 use?
Can I use MiniMax-M2.7 commercially?
How much memory does MiniMax-M2.7 need?
How many parameters does MiniMax-M2.7 have?
Sources
Every figure comes from the developer’s model card, license text and official announcements, and from the model’s Hugging Face repository metadata. Each model page lists its sources and the date they were checked; the register is rechecked every week.