Moonshot's frontier model ranking ECI 145. strong performance across major benchmarks. excellent intelligence. Released Nov 2025. Open weights available.
Metrics
Performance
ECI Score145.1
Epoch Capabilities Index score
ECI Normalized92
Normalized ECI score (0-100)
Training Compute4.2e24 FLOP
Total compute used during training
Info
ProviderMoonshot
Organization that created the model
CountryChina
Model development country
Model AccessibilityOpen weights (restricted use)
Access method (API, download, etc.)
Open WeightsYes
Model weights publicly available
Epoch Model IDkimi-k2-thinking
Epoch AI benchmark identifier
Other
Release DateNov 6, 2025
Score Breakdown
code
20%78
Code generation, understanding, and debugging
math
20%83
Mathematical reasoning and problem solving
reasoning
15%87
Multi-step logical reasoning
intelligence
30%92
Overall reasoning and task completion ability
instruction following
83
Compatibility(5)
Sources(1)
Developed by
Moonshot AIBenchmark Performance
Compare ModelsEpoch Capabilities Index
145.1/ 155
| Dimension | Score |
|---|---|
| Intelligence | 92.0 |
| Reasoning | 87.0 |
| Math | 83.0 |
| Code | 78.0 |
| Instruction Following | 83.0 |
| Overall | 92 |
Loading chart...
Data: Epoch AI (CC-BY 4.0)