ai.paski.devMiniMax M1
llmBilled as the first open-weight large-scale hybrid-attention reasoning model, combining MoE with lightning attention for a native 1M-token context.
Metadata
| org | MiniMax |
| released | 2025-06-17 |
| params | 456B-A45.9B |
| license | apache-2.0 |
| kind | llm |
| hf id | MiniMaxAI/MiniMax-M1-80k |
Benchmarks
| MMLU-Pro | 81.1 | source ↗ |
| GPQA | 70 | source ↗ |
| AIME 2025 | 76.9 | source ↗ |
| SWE-bench | 56 | source ↗ |
Lineage
No recorded lineage — root or standalone model.
view in full graph →