MiniMax-AI/MiniMax-M1
MiniMax-M1 is an open-weight large-scale hybrid-attention reasoning model released by MiniMax.

Velocity · 7d
+8.8
★ / day
Trend
→steady
star history
MiniMax-M1 claims to be the world’s first open-weight large-scale hybrid-attention reasoning model. The repository provides model weights and inference code for this reasoning model. It includes evaluation benchmarks, training methodology documentation, and is designed for research and deployment purposes in the reasoning model space.