ai.paski.devMiniMax M2
llmA 230B-parameter (10B active) MoE model that deliberately returned to full attention over M1's linear attention, optimized for coding and agentic workflows.
Metadata
| org | MiniMax |
| released | 2025-10-27 |
| params | 230B-A10B |
| license | modified-mit |
| kind | llm |
| hf id | MiniMaxAI/MiniMax-M2 |
Benchmarks
| MMLU-Pro | 82 | source ↗ |
| GPQA | 78 | source ↗ |
| AIME 2025 | 78 | source ↗ |
| SWE-bench | 69.4 | source ↗ |
Lineage
↓ finetuneMiniMax M2.1