GPT-6 Astra scores 70.5 on the Respan Index (#2), ahead of Grok 4.7 at 49.8 (#23). GPT-6 Astra leads on 23 of the 23 benchmarks both report. Grok 4.7 is 6.7x cheaper per token at list price. GPT-6 Astra has the larger context window (1.1M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| FrontierSWE V2 (Proximal Labs) |
| 65.5% |
| 29.5% |
| Furniture Assembly (Epoch AI) | 80% | 20.8% |
| GPQA Diamond | 96% | 92.7% |
| LiveBench Agentic Coding (LiveBench) | 57.3% | 54% |
| LiveBench Coding (LiveBench) | 80.4% | 77.2% |
| LiveBench Data Analysis (LiveBench) | 83% | 76.9% |
| LiveBench Instruction Following (LiveBench) | 75.6% | 75.3% |
| LiveBench Language (LiveBench) | 89.4% | 80.1% |
| LiveBench Mathematics (LiveBench) | 96.8% | 95.7% |
| LiveBench Reasoning (LiveBench) | 92.7% | 82.7% |
| Mystery Game Puzzles (Epoch AI) | 84% | 29% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 100% | 98.1% |
| SAGE (Vals AI) | 46.4% | 40.8% |
| SimpleQA Verified | 75.6% | 56% |
| Terminal-Bench 4.0 (Vals AI) | 59.6% | 28.8% |
| Vending-Bench 2 (Andon Labs) | 15514.7 | 10536.83 |
| LMArena Agent (LMArena) | 0.1227 | 0.0404 |
| LMArena Elo | 1477 | 1443 |
| LMArena WebDev (LMArena) | 1788 | 1638 |
Higher is better. Each score links to where it was published.