GPT-5.6 Terra scores 53.5 on the Respan Index (#14), ahead of GPT-5.5 at 52.5 (#15). GPT-5.5 leads on 15 of the 25 benchmarks both report. GPT-5.6 Terra is 2.5x 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
| 85.3% |
| 86% |
| Furniture Assembly (Epoch AI) | 44.2% | 54.2% |
| GPQA Diamond | 90.7% | 92.9% |
| LiveBench Agentic Coding (LiveBench) | 54% | 54.9% |
| LiveBench Coding (LiveBench) | 82.1% | 78.2% |
| LiveBench Data Analysis (LiveBench) | 81.6% | 79.3% |
| LiveBench Instruction Following (LiveBench) | 70.7% | 64.6% |
| LiveBench Language (LiveBench) | 87.4% | 82.9% |
| LiveBench Mathematics (LiveBench) | 95.9% | 94.9% |
| LiveBench Reasoning (LiveBench) | 89.7% | 90.6% |
| Mystery Game Puzzles (Epoch AI) | 56% | 35% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 84.4% | 99.7% |
| SAGE (Vals AI) | 51.5% | 47% |
| SimpleQA Verified | 63% | 43.2% |
| Vending-Bench 2 (Andon Labs) | 7523.84 | 7343.21 |
| WeirdML (Håvard Tveit Ihle) | 84.9% | 78.3% |
| LMArena Agent (LMArena) | 0.0415 | 0.004 |
| LMArena Elo | 1466 | 1466 |
| LMArena Vision (LMArena) | 1287 | 1268 |
| SimpleBench (SimpleBench) | 69% | 48.9% |
| SWE-Bench Pro | 58.6% | 63.4% |
Higher is better. Each score links to where it was published.