Kimi Main
Based on the current Arena #2 preview average score.
Compare Chinese generalist agents for reading, writing, support, and enterprise workflow fit.
Use case: Chinese enterprise assistants and document-heavy workflows
Based on the current Arena #2 preview average score.
Sorted by critical-failure rate, not a universal safety guarantee.
Prioritizes cost tier, then score.
| Metric | Kimi Main | GLM Main |
|---|---|---|
| Overall | 82 | 81 |
| Pass rate | 87% | 88% |
| Critical | 12% | 7% |
| Format pass | 100% | 100% |
| Win rate | 5% | 5% |
| Cost tier | standard | standard |