SGI-Bench Dry Experiment: leaderboard
SGI-Bench subscore for dry-lab scientific tasks such as analysis, simulation planning, and computational experiment design.
Metric: Dry Experiment Score. Source: github.com. Saturation forecast: Around 2030. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 36.64 | #80 |
| 2 | Claude Sonnet 4.5 | 35.79 | #142 |
| 3 | O4 Mini | 35.79 | #180 |
| 4 | Claude Opus 4.1 | 34.69 | #140 |
| 5 | GPT-4.1 | 34.32 | #248 |
| 6 | Grok 4 | 33.71 | #181 |
| 7 | Qwen 3 Max | 33.21 | #208 |
| 8 | O3 | 31.73 | #124 |
| 9 | GPT-5.1 | 31 | #135 |
| 10 | GPT-5 | 29.89 | #98 |
| 11 | Intern-S1 | 28.79 | #285 |
| 12 | Qwen 3 VL 235B A22B | 28.41 | #238 |
| 13 | GPT-5.2 Pro | 28.04 | #61 |
| 14 | GPT-4o | 26.94 | #343 |
| 15 | Gemini 2.5 Pro | 22.51 | #152 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sgi-bench-dry-experiment · How It Works · Data refreshed daily, snapshot 2026-10-20.