GPT-5.2 scores 43.5 on the Respan Index (#36), ahead of Grok 4.5 at 43.2 (#38). Grok 4.5 leads on 13 of the 22 benchmarks both report. Grok 4.5 is 1.6x cheaper per token at list price. Grok 4.5 has the larger context window (500K tokens).
* Estimated: no published scores in that category. How the Respan Index works
| 92.4% |
| 93.4% |
| LiveBench Agentic Coding (LiveBench) | 50.3% | 56.5% |
| LiveBench Coding (LiveBench) | 76.1% | 68.6% |
| LiveBench Data Analysis (LiveBench) | 78.2% | 73% |
| LiveBench Instruction Following (LiveBench) | 61.8% | 71.5% |
| LiveBench Language (LiveBench) | 79.8% | 82.8% |
| LiveBench Mathematics (LiveBench) | 93.2% | 90.8% |
| LiveBench Reasoning (LiveBench) | 83.2% | 87.2% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 96.1% | 97.8% |
| SAGE (Vals AI) | 49.3% | 35% |
| SimpleQA Verified | 37.1% | 48.3% |
| Vending-Bench 2 (Andon Labs) | 3591.33 | 3887.43 |
| WeirdML (Håvard Tveit Ihle) | 72.2% | 46.4% |
| LMArena Elo | 1412 | 1450 |
| LMArena Vision (LMArena) | 1243 | 1279 |
| LMArena WebDev (LMArena) | 1416 | 1552 |
| tau2-bench Banking Knowledge (Sierra) | 32.2% | 47.9% |
| SimpleBench (SimpleBench) | 45.8% | 70% |
Higher is better. Each score links to where it was published.