Claude Opus 5 scores 63.9 on the Respan Index (#7), ahead of Claude Fable 5 at 62.2 (#9). Claude Fable 5 leads on 16 of the 31 benchmarks both report. Claude Opus 5 is 2.0x cheaper per token at list price.
* Estimated: no published scores in that category. How the Respan Index works
| 90.2% |
| 73.2% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 87% | 85.6% |
| FrontierSWE V2 (Proximal Labs) | 47% | 52% |
| Furniture Assembly (Epoch AI) | 35.8% | 60.8% |
| GPQA Diamond | 85.9% | 93.9% |
| LiveBench Agentic Coding (LiveBench) | 62.2% | 65.2% |
| LiveBench Coding (LiveBench) | 86% | 81.4% |
| LiveBench Data Analysis (LiveBench) | 80.5% | 74.6% |
| LiveBench Instruction Following (LiveBench) | 75.8% | 63.8% |
| LiveBench Language (LiveBench) | 90.7% | 88.7% |
| LiveBench Mathematics (LiveBench) | 96% | 95.7% |
| LiveBench Reasoning (LiveBench) | 89.7% | 91.2% |
| Mystery Game Puzzles (Epoch AI) | 52% | 59% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 100% | 98.9% |
| SAGE (Vals AI) | 51.9% | 49.4% |
| SimpleQA Verified | 70.7% | 59.9% |
| SWE-rebench 2026-05-15 to 2026-07-01 (Nebius) | 64.5% | 63.4% |
| Terminal-Bench 4.0 (Vals AI) | 41.4% | 53.5% |
| Vending-Bench 2 (Andon Labs) | 5680.26 | 11181.87 |
| WeirdML (Håvard Tveit Ihle) | 91.9% | 91.8% |
| LMArena Agent (LMArena) | 0.0821 | 0.0867 |
| LMArena Elo | 1504 | 1490 |
| LMArena Vision (LMArena) | 1309 | 1289 |
| LMArena WebDev (LMArena) | 1625 | 1695 |
| tau2-bench Banking Knowledge (Sierra) | 39.7% | 48.7% |
| OSWorld-Verified (XLANG) | 86% | 83.4% |
| SimpleBench (SimpleBench) | 81.9% | 80.6% |
Higher is better. Each score links to where it was published.