SciVQR: leaderboard
Multimodal scientific reasoning on SciVQR (2026), 3,254 image-based scientific questions from competitions, entrance exams and graduate-level sources in English and Chinese, answered with zero-shot chain-of-thought, across mathematics, physics, chemistry, biology, geography and astronomy. Overall accuracy in percent, the mean of the six subject accuracies; higher is better.
Metric: Overall accuracy (%) with zero-shot chain-of-thought prompting on SciVQR (3,254 image-based scientific questions from competitions, entrance exams and graduate-level sources, 78 percent multiple-choice and 22 percent free-form, in English and Chinese, high-school to graduate level); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 15 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | O4 Mini | 79.3 | #368 | |
| 2 | Gemini 2.5 Flash (Preview 04-17) | 79.3 | #453 | |
| 3 | O1 | 77.2 | #403 | |
| 4 | MiMo-VL-7B-RL-2508 | 69.2 | #823 | |
| 5 | Qwen 2.5 VL 72B Instruct | 61.8 | #841 | |
| 6 | GPT-4o | 60.7 | #751 | |
| 7 | Keye-VL-1.5-8B | Thinking | 59.7 | #1720 |
| 8 | InternVL3-78B-Instruct | 53.6 | #1415 | |
| 9 | InternVL3-14B-Instruct | 49.9 | #1801 | |
| 10 | Qwen 2.5 VL 7B Instruct | 45 | #1902 | |
| 11 | InternVL3-8B-Instruct | 43.6 | #2872 | |
| 12 | LLaVA-OneVision-7B | 32.5 | #2981 | |
| 13 | Molmo-7B-D-0924 | 29.8 | #7207 | |
| 14 | MiniCPM-o-2.6 | 28.8 | #2423 | |
| 15 | LLaVA-NeXT-7B | 19.5 | #5032 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=scivqr · How It Works · Data refreshed daily, snapshot 2026-11-21.