Up to four models
Claude Fable 5 vs Kimi K2.7 Code
On the index (GPT-4 = 100), Claude Fable 5 scores 183 and Kimi K2.7 Code 151. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
Claude Fable 5
183167–199
WeightsClosed
ReleasedJun 2026
Price in / out$10 / $50
Context1M
Core tests taken6 of 9
Kimi K2.7 Code
151135–167
WeightsOpen
ReleasedJun 2026
Price in / out$0.70 / $3.75
Context262K
Core tests taken3 of 9
Claude Fable 5 leads Kimi K2.7 Code by 33 points, ahead on 3 of 3 shared core tests, at 14× the price.
Test by testthird-party results
Claude Fable 5Kimi K2.7 Code
Every result as a table
| Test | Claude Fable 5 | Kimi K2.7 Code |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 93.9% | 87.9% |
| LiveBench Data Analysis | 80.5% | 62.7% |
| LiveBench IF | 75.8% | 56.3% |
| LiveBench Language | 90.7% | 77.9% |
| LiveBench Reasoning | 89.6% | 82.8% |
| SimpleBench | 81.9% | 57.9% |
| SimpleQA Verified | 70.7% | 36.5% |
| Maths | ||
| FrontierMath (tiers 1-3) | 87.0% | 54.0% |
| AIME-style maths (OTIS mock) | 100.0% | 95.6% |
| LiveBench Mathematics | 96.0% | 79.6% |
| Chess puzzles | 47.0% | 14.0% |
| FrontierMath (tier 4) | 90.2% | 12.2% |
| Surface Evolver Bench | 95.0% | 48.8% |
| Coding | ||
| LiveBench Agentic Coding | 62.2% | 45.7% |
| LiveBench Coding | 86.0% | 74.0% |
| Code Migration | 55.1% | 25.4% |
| LiveCodeBench | 89.8% | 82.0% |
| ProgramBench | 2.0% | 0.0% |
| SWE-bench | 95.0% | 78.2% |
| Vibe Code Bench v1.1 | 90.3% | 47.2% |
| DeepSWE | 69.9% | 30.5% |
| WeirdML | 91.9% | 54.1% |
| Agents | ||
| APEX-Agents | 63.6% | 37.6% |
Tests at least two of these models have taken. Bold rows are the core tests.