Claude Opus 5.5 scores 72.4 on the Respan Index (#1), ahead of GPT-6 Astra at 70.5 (#2). GPT-6 Astra leads on 15 of the 27 benchmarks both report. Claude Opus 5.5 is 2.5x cheaper per token at list price. GPT-6 Astra has the larger context window (1.1M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| FrontierSWE V2 (Proximal Labs) |
| 62.3% |
| 65.5% |
| Furniture Assembly (Epoch AI) | 83.3% | 80% |
| GPQA Diamond | 90.6% | 96% |
| LiveBench Agentic Coding (LiveBench) | 71.7% | 57.3% |
| LiveBench Coding (LiveBench) | 89.3% | 80.4% |
| LiveBench Data Analysis (LiveBench) | 80.3% | 83% |
| LiveBench Instruction Following (LiveBench) | 67% | 75.6% |
| LiveBench Language (LiveBench) | 86.3% | 89.4% |
| LiveBench Mathematics (LiveBench) | 97.1% | 96.8% |
| LiveBench Reasoning (LiveBench) | 92.2% | 92.7% |
| MirrorCode (Epoch AI) | 77.4% | 46.7% |
| Mystery Game Puzzles (Epoch AI) | 71% | 84% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 100% | 100% |
| SAGE (Vals AI) | 45.8% | 46.4% |
| SimpleQA Verified | 72.2% | 75.6% |
| Terminal-Bench 4.0 (Vals AI) | 65.2% | 59.6% |
| Vending-Bench 2 (Andon Labs) | 9235.25 | 15514.7 |
| LMArena Agent (LMArena) | 0.1382 | 0.1227 |
| LMArena Elo | 1504 | 1477 |
| LMArena WebDev (LMArena) | 1815 | 1788 |
| SimpleBench (SimpleBench) | 88.4% | 86.5% |
| AutomationBench 1.0.6 | 40% | 41.4% |
| Terminal-Bench Science 0.1 | 58.7% | 64.6% |
Higher is better. Each score links to where it was published.