Kimi K1.5: benchmark results
Provider: Moonshot. Released 2025-01-20. Access: API.
Aggregate Rating 1600 ± 26, rank #1016 of 10916 rated models, from 13 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LLM Stats (MathVista) | 74.9 | Score (%) | 81.6 |
| SuperCLUE General (May 2025) - Math Reasoning | 59.68 | Score | 75.6 |
| SuperCLUE General (May 2025) - Agent | 50.68 | Score | 56.4 |
| ZeroEval MATH-500 | 96.2 | MATH-500 Score | 54.7 |
| SuperCLUE General (May 2025) - Precise Instruction Following | 25.77 | Score | 53.8 |
| SuperCLUE General (May 2025) - Overall | 53.72 | Score | 51.3 |
| MathVision | 38.6 | Overall Accuracy (%) | 51 |
| LLM Stats (C-Eval) | 88.3 | Score (%) | 47.4 |
| SuperCLUE General (May 2025) - Science Reasoning | 38.61 | Score | 44.9 |
| LLM Stats (AIME 2024) | 77.5 | Score (%) | 42 |
| SuperCLUE General (May 2025) - Code Generation | 73.37 | Score | 38.5 |
| LLM Stats Score | 16.61 | LLM Stats Score (conservative rating) | 37.2 |
Interactive version: theaggregate.ai/model?slug=kimi-k1-5 · How It Works · Data refreshed daily, snapshot 2026-11-08.