GPT-5.6 Terra scores 53.5 on the Respan Index (#14), ahead of Grok 4.6 at 51.0 (#19). GPT-5.6 Terra leads on 17 of the 26 benchmarks both report. Grok 4.6 is 1.5x 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
| 86% |
| 66% |
| Furniture Assembly (Epoch AI) | 54.2% | 40% |
| GPQA Diamond | 92.9% | 94% |
| LiveBench Agentic Coding (LiveBench) | 54.9% | 57% |
| LiveBench Coding (LiveBench) | 78.2% | 76.8% |
| LiveBench Data Analysis (LiveBench) | 79.3% | 73.9% |
| LiveBench Instruction Following (LiveBench) | 64.6% | 71.9% |
| LiveBench Language (LiveBench) | 82.9% | 83.7% |
| LiveBench Mathematics (LiveBench) | 94.9% | 92.6% |
| LiveBench Reasoning (LiveBench) | 90.6% | 90.5% |
| Mystery Game Puzzles (Epoch AI) | 35% | 34% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 99.7% | 99.2% |
| SAGE (Vals AI) | 47% | 28.9% |
| SimpleQA Verified | 43.2% | 49.3% |
| Terminal-Bench 4.0 (Vals AI) | 22.7% | 17.2% |
| Vending-Bench 2 (Andon Labs) | 7343.21 | 9047.03 |
| WeirdML (Håvard Tveit Ihle) | 78.3% | 67.3% |
| LMArena Agent (LMArena) | 0.004 | 0.0128 |
| LMArena Elo | 1466 | 1454 |
| LMArena Vision (LMArena) | 1268 | 1264 |
| DeepSWE v1.1 | 69.6% | 65.9% |
| SimpleBench (SimpleBench) | 48.9% | 75.9% |
Higher is better. Each score links to where it was published.