Up to four models
Claude Opus 4.6 vs Kimi K2.5
On the index (GPT-4 = 100), Claude Opus 4.6 scores 165 and Kimi K2.5 145. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
Claude Opus 4.6
165149–181
WeightsClosed
ReleasedFeb 2026
Price in / out$5.0 / $25
Context1M
Core tests taken7 of 9
Kimi K2.5
145130–161
WeightsOpen
ReleasedJan 2026
Price in / out$0.50 / $2.5
Context262K
Core tests taken5 of 9
Claude Opus 4.6 leads Kimi K2.5 by 20 points, ahead on 4 of 4 shared core tests, at 10× the price.
Test by testthird-party results
Claude Opus 4.6Kimi K2.5
Every result as a table
| Test | Claude Opus 4.6 | Kimi K2.5 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 90.5% | 87.6% |
| Humanity's Last Exam | 34.4% | 24.4% |
| GPQA Diamond (Vals AI) | 89.6% | 84.1% |
| MMLU Pro | 89.1% | 85.9% |
| MMMU Pro | 83.9% | 84.3% |
| CL-bench | 20.7% | 19.3% |
| SimpleBench | 67.6% | 46.8% |
| SimpleQA Verified | 47.0% | 34.3% |
| Maths | ||
| AIME-style maths (OTIS mock) | 94.4% | 92.2% |
| FrontierMath (2025 set) | 40.7% | 27.9% |
| FrontierMath tier 4 (2025 set) | 22.9% | 4.2% |
| Coding | ||
| SWE-bench Verified | 78.7% | 73.8% |
| Terminal-Bench 2.0 (Terminus 2 runs) | 62.9% | 43.2% |
| LiveCodeBench | 84.7% | 83.9% |
| SWE-bench | 78.2% | 70.0% |
| Vibe Code Bench v1.1 | 57.6% | 17.5% |
| WeirdML | 78.0% | 45.6% |
| Professional work | ||
| CorpFin v2 | 67.0% | 68.3% |
| MedCode | 49.1% | 39.3% |
| MedScribe | 86.7% | 76.4% |
| MortgageTax | 68.5% | 66.5% |
| TaxEval v2 | 76.0% | 74.2% |
| Novel problems | ||
| ARC-AGI-2 | 69.2% | 11.8% |
| ARC-AGI-1 | 94.0% | 65.3% |
Task length · METR47 sec
Tests at least two of these models have taken. Bold rows are the core tests.