Claude Opus 5.5 scores 72.4 on the Respan Index (#1), ahead of Kimi K3 at 50.6 (#21). Claude Opus 5.5 leads on 21 of the 24 benchmarks both report. Kimi K3 is 1.3x cheaper per token at list price. Kimi K3 has the larger context window (1.0M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| Furniture Assembly (Epoch AI) |
| 83.3% |
| 34.2% |
| GPQA Diamond | 90.6% | 93.5% |
| LiveBench Agentic Coding (LiveBench) | 71.7% | 62.2% |
| LiveBench Coding (LiveBench) | 89.3% | 81.4% |
| LiveBench Data Analysis (LiveBench) | 80.3% | 78.7% |
| LiveBench Instruction Following (LiveBench) | 67% | 71.4% |
| LiveBench Language (LiveBench) | 86.3% | 85.5% |
| LiveBench Mathematics (LiveBench) | 97.1% | 84.4% |
| LiveBench Reasoning (LiveBench) | 92.2% | 90.7% |
| Mystery Game Puzzles (Epoch AI) | 71% | 26% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 100% | 97.2% |
| SAGE (Vals AI) | 45.8% | 52.8% |
| SimpleQA Verified | 72.2% | 50.6% |
| Terminal-Bench 4.0 (Vals AI) | 65.2% | 17.2% |
| Vending-Bench 2 (Andon Labs) | 9235.25 | 5165.04 |
| LMArena Agent (LMArena) | 0.1382 | 0.0418 |
| LMArena Elo | 1504 | 1488 |
| LMArena WebDev (LMArena) | 1815 | 1658 |
| SimpleBench (SimpleBench) | 88.4% | 60.7% |
| Humanity's Last Exam (with tools) | 67.7% | 56% |
Higher is better. Each score links to where it was published.