KVBench (Brief Caption) - Physics (Chinese): leaderboard

KVBench (2026) tests whether text-to-image models can draw correct senior high-school physics content, such as diagrams and scientific illustrations, from Chinese prompts given as a brief caption, so the model must supply the textbook knowledge itself. Qwen2.5-VL-32B checks each image against a list of textbook-grounded constraints; the score is the percentage satisfied, higher is better.

Metric: Checklist score (%) on the physics prompts in Chinese, each given as a brief caption, so the model must supply the textbook knowledge itself: KVBench knowledge-intensive text-to-image prompts from over 30 senior high-school textbooks (150 per subject and language, each with a textbook reference image); Qwen2.5-VL-32B-Instruct answers a checklist of four to six binary questions (key objects, attributes, spatial relations, reasoning outcomes) about each generated image, and the score is the mean share of satisfied items, in percent; higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 14 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Nano Banana Pro (Gemini 3 Pro Image)51.58
2Seedream 4.040.27
3GPT-Image (KVBench checkpoint unspecified)39.1
4FLUX.2-dev36.4
5FLUX 2 Max27.97
6OmniGen226.87#4579
7FLUX.1-dev26.8
8BAGEL-7B-MoT22.78#4619
9Qwen-Image20.53
10Z-Image13.43
11Show-o213.33
12FLUX.1-Krea-dev10.57
13SD3.5-Large (KVBench GGUF build, quantization unspecified)6
14Janus-Pro5.08

Interactive version: theaggregate.ai/benchmark?slug=kvbench-brief-caption-physics-chinese · How It Works · Data refreshed daily, snapshot 2026-11-21.