MiniMax M2.5 230B Architecture
Back to MiniMax M2.5 230B in the gallery
MiniMax
Popular 230B coder that opts for a classic architecture instead of the newer hybrid-attention ideas.
Architecture and specifications
- Scale
- 230B total, 10B active (4.3% active)
- Context (tokens)
- 196,608
- License
- Modified MIT License
- Date
- 2026-02-12
- Decoder type
- Sparse MoE
- Attention
- GQA with QK-Norm
- Layer mix
- 62 GQA
- KV cache / token (bf16) info
- 248 KiB · High
- Key detail
- Deliberately avoids sliding-window or linear-attention hybrids while keeping a 10B active path and three MTP modules.
Training and benchmarks
- Pretraining data
- 29.2T tokens (M2-series base) The series report gives 19.9T tokens in the constant phase and 9.3T in decay and context extension. It does not separate pretraining budgets by M2-series checkpoint. Technical report
- Post-training methods
- SFT, RLVR The M2-series recipe uses SFT and mixed-domain RL with executable task rewards, plus process and completion-time rewards. The recipe evolves across M2, M2.5, and M2.7. Technical report