Gemini 3.7 Flash scores 48.6 on the Respan Index (#26), ahead of GPT-5.4 at 47.2 (#28). Gemini 3.7 Flash leads on 13 of the 22 benchmarks both report. Gemini 3.7 Flash is 3.8x cheaper per token at list price. GPT-5.4 has the larger context window (1.1M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| 71.6% |
| 78.6% |
| Furniture Assembly (Epoch AI) | 26.7% | 37.5% |
| GPQA Diamond | 94.8% | 92.8% |
| LiveBench Agentic Coding (LiveBench) | 58.3% | 53.8% |
| LiveBench Coding (LiveBench) | 78.9% | 77.5% |
| LiveBench Data Analysis (LiveBench) | 68% | 79.3% |
| LiveBench Instruction Following (LiveBench) | 79.9% | 70.2% |
| LiveBench Language (LiveBench) | 85.5% | 82.6% |
| LiveBench Mathematics (LiveBench) | 93.5% | 94.1% |
| LiveBench Reasoning (LiveBench) | 87.8% | 88.1% |
| Mystery Game Puzzles (Epoch AI) | 37% | 37% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 97.2% | 97.8% |
| SAGE (Vals AI) | 49.2% | 43.3% |
| SimpleQA Verified | 69.2% | 45.1% |
| LMArena Agent (LMArena) | -0.0149 | 0.0105 |
| LMArena Elo | 1488 | 1453 |
| LMArena Vision (LMArena) | 1296 | 1284 |
| LMArena WebDev (LMArena) | 1592 | 1395 |
Higher is better. Each score links to where it was published.