Up to four models
Claude Opus 4.7 vs Kimi K2.6
On the index (GPT-4 = 100), Claude Opus 4.7 scores 167 and Kimi K2.6 152. 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 4.7
167151–183
WeightsClosed
ReleasedApr 2026
Price in / out$5.0 / $25
Context1M
Core tests taken6 of 9
Kimi K2.6
152136–169
WeightsOpen
ReleasedApr 2026
Price in / out$0.65 / $3.41
Context262K
Core tests taken4 of 9
Claude Opus 4.7 leads Kimi K2.6 by 15 points, ahead on 3 of 4 shared core tests, at 7.5× the price.
Test by testthird-party results
Claude Opus 4.7Kimi K2.6
Every result as a table
| Test | Claude Opus 4.7 | Kimi K2.6 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 90.1% | 90.8% |
| LiveBench Data Analysis | 78.3% | 65.1% |
| LiveBench IF | 66.7% | 64.4% |
| LiveBench Language | 77.9% | 75.1% |
| LiveBench Reasoning | 87.2% | 79.4% |
| GPQA Diamond (Vals AI) | 90.1% | 89.1% |
| MMLU Pro | 89.9% | 87.6% |
| MMMU Pro | 85.5% | 86.3% |
| EBR-bench | 19.1% | 2.4% |
| SimpleQA Verified | 51.7% | 34.9% |
| Maths | ||
| FrontierMath (tiers 1-3) | 70.2% | 57.2% |
| AIME-style maths (OTIS mock) | 97.8% | 96.1% |
| LiveBench Mathematics | 92.8% | 84.3% |
| Chess puzzles | 41.0% | 20.0% |
| FrontierMath (2025 set) | 43.8% | 39.0% |
| FrontierMath tier 4 (2025 set) | 22.9% | 14.6% |
| FrontierMath (tier 4) | 31.7% | 25.6% |
| Coding | ||
| SWE-bench Verified | 83.5% | 76.6% |
| LiveBench Agentic Coding | 50.7% | 46.9% |
| LiveBench Coding | 82.1% | 78.6% |
| Code Migration | 43.9% | 27.8% |
| LiveCodeBench | 85.1% | 86.8% |
| ProgramBench | 0.0% | 0.0% |
| SWE-bench | 82.0% | 76.2% |
| Vibe Code Bench v1.1 | 71.0% | 37.9% |
| WeirdML | 76.4% | 55.9% |
| Agents | ||
| DTBench | 94.7% | 90.9% |
| Furniture assembly | 33.3% | 21.7% |
| LMCA | 52.2% | 37.3% |
| Mystery game puzzles | 28.0% | 18.0% |
| Professional work | ||
| Finance Agent (v2) | 51.5% | 44.9% |
| CorpFin v2 | 66.1% | 66.7% |
| EMB | 63.7% | 57.9% |
| LegalBench | 85.3% | 84.7% |
| Legal Research Bench | 38.5% | 15.9% |
| MedCode | 54.9% | 40.1% |
| MedScribe | 83.0% | 78.1% |
| MortgageTax | 70.3% | 65.8% |
| Tax Agent Bench | 23.7% | 12.4% |
| TaxEval v2 | 75.3% | 74.7% |
Tests at least two of these models have taken. Bold rows are the core tests.