GPT-5.6 Terra scores 53.5 on the Respan Index (#14), ahead of Claude Opus 4.8 at 52.2 (#16). GPT-5.6 Terra leads on 14 of the 26 benchmarks both report. GPT-5.6 Terra is 2.2x cheaper per token at list price. GPT-5.6 Terra has the larger context window (1.1M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| FrontierMath Tiers 1-3 v2 (Epoch AI) |
| 80% |
| 86% |
| Furniture Assembly (Epoch AI) | 42.5% | 54.2% |
| GPQA Diamond | 93.6% | 92.9% |
| LiveBench Agentic Coding (LiveBench) | 50.5% | 54.9% |
| LiveBench Coding (LiveBench) | 81.8% | 78.2% |
| LiveBench Data Analysis (LiveBench) | 66% | 79.3% |
| LiveBench Instruction Following (LiveBench) | 72% | 64.6% |
| LiveBench Language (LiveBench) | 79.7% | 82.9% |
| LiveBench Mathematics (LiveBench) | 94.3% | 94.9% |
| LiveBench Reasoning (LiveBench) | 89.2% | 90.6% |
| Mystery Game Puzzles (Epoch AI) | 36% | 35% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 98.3% | 99.7% |
| SAGE (Vals AI) | 54.8% | 47% |
| SimpleQA Verified | 53% | 43.2% |
| Terminal-Bench 4.0 (Vals AI) | 23.2% | 22.7% |
| Vending-Bench 2 (Andon Labs) | 5787.43 | 7343.21 |
| WeirdML (Håvard Tveit Ihle) | 82.9% | 78.3% |
| LMArena Agent (LMArena) | 0.0664 | 0.004 |
| LMArena Elo | 1453 | 1466 |
| LMArena Vision (LMArena) | 1286 | 1268 |
| SimpleBench (SimpleBench) | 64.8% | 48.9% |
| SWE-Bench Pro | 69.2% | 63.4% |
Higher is better. Each score links to where it was published.