MiniMax-M2.7
MiniMax's self-evolving agent model pioneering recursive self-improvement with frontier agentic coding performance at a fraction of competitor cost
MiniMax-M2.7
MiniMax • March 2026
Training Data
Up to early 2026
MiniMax-M2.7
March 2026
Parameters
230 billion (10B active)
Training Method
Mixture of Experts with RL and autonomous self-evolution loops
Context Window
204,800 tokens
Knowledge Cutoff
Not disclosed
Key Features
Self-Evolving Agent Model • Agent Teams & Dynamic Tool Search • Open Weights (205K Context)
Capabilities
Coding: Outstanding
Agent Tasks: Outstanding
Reasoning: Excellent
What's New in This Version
First model with autonomous self-evolution capabilities, achieving 30% internal performance improvement through 100+ optimization rounds and 97% skill adherence vs M2.5
MiniMax's self-evolving agent model pioneering recursive self-improvement with frontier agentic coding performance at a fraction of competitor cost
What's New in This Version
First model with autonomous self-evolution capabilities, achieving 30% internal performance improvement through 100+ optimization rounds and 97% skill adherence vs M2.5
Technical Specifications
Key Features
Capabilities
Other MiniMax Models
Explore more models from MiniMax
MiniMax-M2.5
MiniMax's flagship model matching frontier performance at 1/20th the cost with 80.2% SWE-bench Verified
MiniMax-M2.5-Lightning
Ultra-fast variant of M2.5 generating 100 tokens per second at $1/hour continuous operation
MiniMax-M2.1
Lightweight coding-focused model with strong multi-language programming and agentic capabilities
