← All models

MiniMax M2.7

MiniMax

flagship

Capabilities

VisionToolsCodeLong ContextExtended ThinkingStructured OutputMultilingual

Strengths

  • Self-evolving model — autonomous scaffold optimization
  • SWE-bench Verified 78% (near Opus level)
  • 230B total / 10B active MoE — very efficient
  • 1M context
  • Open-sourced April 2026
  • Strong engineering and agentic tasks

Weaknesses

  • Newer provider
  • China-based

Pricing

Input / 1M tokens

$0.24

Output / 1M tokens

$0.96

Context window

1,000k

Transparency

Open weights

7.0 / 10

Open training data

3.0 / 10

Open methodology

5.0 / 10

Licence openness

7.0 / 10

Provider disclosure

2.0 / 10

FMTI company score

N/A

Composite:4.8 / 10

Open weights. Not FMTI-scored.

Sustainability

Inference energy

6.0 / 10

Training footprint

N/A

Provider infrastructure

2.0 / 10

Composite:4.0 / 10

Larger model. No provider energy data.

Task Fitness

question answering
79%
code
86%
mathematics
84%
business communication
64%
analysis
82%
extraction
75%
long-form generation
75%
research
57%
reasoning
82%
summarisation
75%
translation
70%
conversation
75%