Claude Fable 5 scores 62.2 on the Respan Index (#9), ahead of GPT-5.5 at 52.5 (#15). Claude Fable 5 leads on 23 of the 30 benchmarks both report. GPT-5.5 is 1.8x cheaper per token at list price.
* Estimated: no published scores in that category. How the Respan Index works
| 90.2% |
| 72.5% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 87% | 85.3% |
| Furniture Assembly (Epoch AI) | 35.8% | 44.2% |
| GPQA Diamond | 85.9% | 90.7% |
| LiveBench Agentic Coding (LiveBench) | 62.2% | 54% |
| LiveBench Coding (LiveBench) | 86% | 82.1% |
| LiveBench Data Analysis (LiveBench) | 80.5% | 81.6% |
| LiveBench Instruction Following (LiveBench) | 75.8% | 70.7% |
| LiveBench Language (LiveBench) | 90.7% | 87.4% |
| LiveBench Mathematics (LiveBench) | 96% | 95.9% |
| LiveBench Reasoning (LiveBench) | 89.7% | 89.7% |
| MirrorCode (Epoch AI) | 63.9% | 10% |
| Mystery Game Puzzles (Epoch AI) | 52% | 56% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 100% | 84.4% |
| SAGE (Vals AI) | 51.9% | 51.5% |
| SimpleQA Verified | 70.7% | 63% |
| Vending-Bench 2 (Andon Labs) | 5680.26 | 7523.84 |
| WeirdML (Håvard Tveit Ihle) | 91.9% | 84.9% |
| LMArena Agent (LMArena) | 0.0821 | 0.0415 |
| LMArena Elo | 1504 | 1466 |
| LMArena Vision (LMArena) | 1309 | 1287 |
| GSO Opt@1 (GSO) | 76.5% | 37.3% |
| tau2-bench Banking Knowledge (Sierra) | 39.7% | 44.6% |
| SimpleBench (SimpleBench) | 81.9% | 69% |
| OSWorld-Verified | 85% | 78.7% |
| SWE-Bench Pro | 80% | 58.6% |
Higher is better. Each score links to where it was published.