Claude Opus 5 scores 63.9 on the Respan Index (#7), ahead of GPT-5.6 Sol at 60.6 (#10). Claude Opus 5 leads on 18 of the 30 benchmarks both report. GPT-5.6 Sol is 1.3x cheaper per token at list price. GPT-5.6 Sol has the larger context window (1.1M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| 73.2% |
| 82.9% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 85.6% | 89.1% |
| FrontierSWE V2 (Proximal Labs) | 52% | 32.2% |
| Furniture Assembly (Epoch AI) | 60.8% | 56.7% |
| GPQA Diamond | 93.9% | 94.6% |
| LiveBench Agentic Coding (LiveBench) | 65.2% | 56.2% |
| LiveBench Coding (LiveBench) | 81.4% | 83.9% |
| LiveBench Data Analysis (LiveBench) | 74.6% | 79.8% |
| LiveBench Instruction Following (LiveBench) | 63.8% | 71.8% |
| LiveBench Language (LiveBench) | 88.7% | 87.7% |
| LiveBench Mathematics (LiveBench) | 95.7% | 96.2% |
| LiveBench Reasoning (LiveBench) | 91.2% | 91.7% |
| Mystery Game Puzzles (Epoch AI) | 59% | 58% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 98.9% | 100% |
| SAGE (Vals AI) | 49.4% | 52.6% |
| SimpleQA Verified | 59.9% | 69.7% |
| SWE-rebench 2026-05-15 to 2026-07-01 (Nebius) | 63.4% | 62.3% |
| Terminal-Bench 4.0 (Vals AI) | 53.5% | 37.9% |
| Vending-Bench 2 (Andon Labs) | 11181.87 | 9619.37 |
| WeirdML (Håvard Tveit Ihle) | 91.8% | 89.4% |
| LMArena Agent (LMArena) | 0.0867 | 0.0645 |
| LMArena Elo | 1490 | 1484 |
| LMArena Vision (LMArena) | 1289 | 1285 |
| BALROG (BALROG) | 63.4% | 60% |
| tau2-bench Banking Knowledge (Sierra) | 48.7% | 46.9% |
| SimpleBench (SimpleBench) | 80.6% | 71.7% |
Higher is better. Each score links to where it was published.