2506.13585v1
arXiv License
Alibaba / Qwen Team
MiniMax-M1:以閃電注意力高效地擴展測試時運算
MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention
MiniMax(通訊:[email protected];共 128 位貢獻者,完整名單見原論文附錄 A)
Reasoning ModelLightning AttentionReinforcement LearningCISPOMoELong Context開源模型
Liang-Hsun Huang 許願