Gemini 3.8 Flash scores 52.2 on the Respan Index (#16), ahead of Grok 4.7 at 49.8 (#23). Gemini 3.8 Flash leads on 15 of the 24 benchmarks both report. Gemini 3.8 Flash is 2.0x cheaper per token at list price. Gemini 3.8 Flash has the larger context window (1.0M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| FrontierSWE V2 (Proximal Labs) |
| 19.6% |
| 29.5% |
| Furniture Assembly (Epoch AI) | 31.7% | 20.8% |
| GPQA Diamond | 95.3% | 92.7% |
| LiveBench Agentic Coding (LiveBench) | 54.2% | 54% |
| LiveBench Coding (LiveBench) | 72.5% | 77.2% |
| LiveBench Data Analysis (LiveBench) | 54% | 76.9% |
| LiveBench Instruction Following (LiveBench) | 81.4% | 75.3% |
| LiveBench Language (LiveBench) | 87.8% | 80.1% |
| LiveBench Mathematics (LiveBench) | 91.6% | 95.7% |
| LiveBench Reasoning (LiveBench) | 89.3% | 82.7% |
| Mystery Game Puzzles (Epoch AI) | 47% | 29% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 98.9% | 98.1% |
| SAGE (Vals AI) | 35.1% | 40.8% |
| SimpleQA Verified | 69.7% | 56% |
| Terminal-Bench 4.0 (Vals AI) | 19.2% | 28.8% |
| Vending-Bench 2 (Andon Labs) | 5093.79 | 10536.83 |
| LMArena Agent (LMArena) | 0.0297 | 0.0404 |
| LMArena Elo | 1495 | 1443 |
| LMArena WebDev (LMArena) | 1583 | 1638 |
| DeepSWE v1.1 | 73.7% | 71% |
Higher is better. Each score links to where it was published.