MiniMax M1
MiniMax-M1 is a large-scale, open-weight reasoning model designed for extended context and high-efficiency inference. It leverages a hybrid Mixture-of-Experts (MoE) architecture paired with a custom "lightning attention" mechanism, allowing it to process long sequences—up to 1 million tokens—while maintaining competitive FLOP efficiency.
Input
Output
Context
1000K
Max Output
40K
Parameters
—
Technical Specifications
Model TypeChat
Context Window1.0M
Max Output Tokens40K
Parameters—
Training Cutoff—
License—
Capabilities
Input Modalities
text
Output Modalities
text
Features
function_callingjson_modestreamingtext_generation
Benchmarks
#46
Rank (Text)
+5/-5
arena text ci 95
1354
Arena Score (Text)
17,524
Votes (Text)
Resources & Links
Related Tools
Data sourced from official provider APIs and documentation
Last updated: May 5, 2026
Start building with the right model.
From model selection to production, one platform, no fragmentation.