SciEval - Materials Science: leaderboard
SciEval multimodal scientific-reasoning leaderboard — materials-science-discipline accuracy.
Metric: Materials Science (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O3 | 93.85 | #124 |
| 2 | GPT-5 | 93.54 | #98 |
| 3 | Gemini 3 Pro | 91.23 | #80 |
| 4 | Gemini 2.5 Pro | 83.23 | #152 |
| 5 | Claude Sonnet 4.5 | 83.23 | #142 |
| 6 | Qwen 3 VL 235B A22B | 81.85 | #238 |
| 7 | Intern-S1 | 81.69 | #285 |
| 8 | Llama 4 Maverick | 80.77 | #464 |
| 9 | Kimi K2 | 71.38 | #243 |
| 10 | Qwen 3 Max | 69.54 | #208 |
| 11 | Ling-flash-2.0 | 63.69 | #550 |
| 12 | Gemini 2.5 Flash | 62.15 | #245 |
| 13 | GPT-4o | 61.54 | #343 |
| 14 | GPT-5.1 | 61.08 | #135 |
| 15 | GLM-4.5V | 57.08 | #346 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=scieval-materials-science · How It Works · Data refreshed daily, snapshot 2026-10-20.