Moonshot's top-tier model ranking ECI 140. competitive on reasoning and coding tasks. strong intelligence. Released Jul 2025. Open weights available.
Metrics
Performance
ECI Score140.484
Epoch Capabilities Index score
ECI Normalized88
Normalized ECI score (0-100)
Training Compute3.0e24 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-Instruct
Epoch AI benchmark identifier
Other
Release DateJul 12, 2025
Score Breakdown
code
20%75
Code generation, understanding, and debugging
math
20%79
Mathematical reasoning and problem solving
reasoning
15%84
Multi-step logical reasoning
intelligence
30%88
Overall reasoning and task completion ability
instruction following
79
Compatibility(5)
Sources(1)
Developed by
Moonshot AIBenchmark Performance
Compare ModelsEpoch Capabilities Index
140.5/ 155
| Dimension | Score |
|---|---|
| Intelligence | 88.0 |
| Reasoning | 84.0 |
| Math | 79.0 |
| Code | 75.0 |
| Instruction Following | 79.0 |
| Overall | 88 |
Loading chart...
Data: Epoch AI (CC-BY 4.0)