SciVQR - Biology: leaderboard

The biology questions of 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. Accuracy in percent; higher is better.

Metric: Biology 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

#ModelSettingScoreOverall rank
1O4 Mini80.8#368
2O180.6#403
3Gemini 2.5 Flash (Preview 04-17)78.6#453
4GPT-4o72.2#751
5MiMo-VL-7B-RL-250871.1#823
6Qwen 2.5 VL 72B Instruct65.3#841
7Keye-VL-1.5-8BThinking62.2#1720
8InternVL3-78B-Instruct60.6#1415
9InternVL3-14B-Instruct57.2#1801
10Qwen 2.5 VL 7B Instruct55.8#1902
11InternVL3-8B-Instruct50.3#2872
12LLaVA-OneVision-7B40.6#2981
13Molmo-7B-D-092438.1#7207
14MiniCPM-o-2.635.3#2423
15LLaVA-NeXT-7B28.9#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-biology · How It Works · Data refreshed daily, snapshot 2026-11-21.