Back to MiniMax M2 230B in the gallery

MiniMax

MiniMax's flagship returns to full attention and looks like a leaner, sparser cousin of Qwen3.

MiniMax M2 230B architecture diagram
MiniMax M2 230B architecture diagram. Select the diagram to view it at full size.

Architecture and specifications

Scale
230B total, 10B active (4.3% active)
Context (tokens)
196,608
Vocabulary
200,064
License
Modified MIT License
Date
2025-10-23
Decoder type
Sparse MoE
Attention
GQA with QK-Norm and partial RoPE
Layer mix
62 GQA
KV cache / token (bf16) info
248 KiB · High
Key detail
Applies RMSNorm across all 6,144 query and 1,024 key dimensions before splitting them into heads. Each token activates 8 of 256 routed experts.

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
AA Intelligence Index v4.3.2 info
Total score 18.6 General N/A Scientific 7.3 Coding N/A Agents N/A

Further reading and tools