Kimi K2.6 (High): benchmark results
Provider: Moonshot. Released 2026-04-20. Access: Open.
Unified ELO 1665 ± 1, rank #125 of 2032 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenCompass LLM - Math | 75.7 | Score (%) | 85.7 |
| OpenCompass Math - College | 82.7 | Score (%) | 85.7 |
| ATOBench - Basket (Native) | 60 | Grounded verification rate (%; the resource-ownership contra | 75 |
| OpenCompass Reasoning - Common | 78.1 | Score (%) | 75 |
| OpenCompass Knowledge - Humanities | 92.5 | Score (%) | 67.9 |
| SlopCodeBench | 18.88 | Isolated Solved (%) | 62.5 |
| OpenCompass Knowledge - Engineering | 90.8 | Score (%) | 60.7 |
| OpenCompass Math - Competition | 68.6 | Score (%) | 60.7 |
| OpenCompass Knowledge - Common Sense | 89.6 | Score (%) | 57.1 |
| ATOBench - JWT (Native) | 86.7 | Grounded verification rate (%; the reusable-artifact contrac | 50 |
| OpenCompass LLM - Knowledge | 89.8 | Score (%) | 50 |
| OpenCompass LLM - Reasoning | 62.9 | Score (%) | 50 |
Interactive version: theaggregate.ai/model?slug=kimi-k2-6-high · How It Works · Data refreshed daily, snapshot 2026-09-26.