CSVQA: leaderboard
Each vision-language model answers all 1,378 Chinese high-school biology, chemistry, math and physics questions built around an image (1,278 multiple choice, some with more than one correct option, and 100 open) with step-by-step reasoning; the score is accuracy, with a GPT model reading each rationale to pick the answer and GPT-4o scoring open answers. Unlike English science image sets such as ScienceQA, the questions come from Chinese textbooks and exams and most need a calculation.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 27 models tracked. Updated 17 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O1 (2024-12-17) | 49.6 | #314 |
| 2 | Qwen 2.5 VL 72B Instruct | 38.5 | #849 |
| 3 | InternVL3-78B | 37.4 | #755 |
| 4 | Claude 3.7 Sonnet (20250219) | 36.6 | #421 |
| 5 | QVQ-72B-Preview | 36.6 | #1173 |
| 6 | InternVL3-38B | 33.4 | #927 |
| 7 | InternVL3-14B | 31.9 | #1191 |
| 8 | Qwen 2.5 VL 32B Instruct | 28.9 | #1074 |
| 9 | InternVL2.5-78B | 28.4 | #1036 |
| 10 | InternVL2.5-38B | 25.2 | #1366 |
| 11 | Qwen 2 VL 72B Instruct | 24.1 | #1382 |
| 12 | Gemma 3 27B (IT) | 22.9 | #1339 |
| 13 | InternVL2.5-26B | 22.1 | #2191 |
| 14 | InternVL2.5-8B | 20.2 | #2401 |
| 15 | Qwen 2.5 VL 7B Instruct | 17.5 | #1893 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=csvqa · How It Works · Data refreshed daily, snapshot 2026-11-17.