Kimi K3: benchmark results
Provider: Moonshot. Released 2026-07-17. Access: Open.
Unified ELO 1754 ± 1, rank #15 of 1392 rated models, from 258 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Design Arena (3D) | 1435 | Elo | 100 |
| Design Arena (UI Components) | 1378 | Elo | 100 |
| Design Arena (Website) | 1360 | Elo | 100 |
| E-commerce Last Exam | 56.74 | Mean Reward (%) | 100 |
| E-commerce Last Exam - Travel | 48.11 | Mean Reward (%) | 100 |
| Humanity's Last Exam (Self-Reported, No Tools) | 56 | Accuracy (%) | 100 |
| LLM Stats (DeepSearchQA) | 95 | Score (%) | 100 |
| LLM Stats (MathVision) | 97.8 | Score (%) | 100 |
| LLM Stats (Program Bench) | 77.8 | Score (%) | 100 |
| LLM Stats (ZEROBench) | 41 | Score (%) | 100 |
| OpenClawProBench | 81.6 | Overall Score (%) | 100 |
| QuantCodeEval (OpenCode) | 26 | Full Pass Rate (%) | 100 |
Interactive version: theaggregate.ai/model?slug=kimi-k3 · How It Works · Data refreshed daily, snapshot 2026-09-05.