KVBench (Brief Caption) - Chemistry (English): leaderboard

KVBench (2026) tests whether text-to-image models can draw correct senior high-school chemistry content, such as diagrams and scientific illustrations, from English 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 chemistry prompts in English, 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
1GPT-Image (KVBench checkpoint unspecified)51.6
2Nano Banana Pro (Gemini 3 Pro Image)50.4
3Seedream 4.049
4FLUX 2 Max45.2
5FLUX.1-dev41.13
6FLUX.2-dev38.73
7SD3.5-Large (KVBench GGUF build, quantization unspecified)36.53
8Janus-Pro27.07
9Show-o226.67
10BAGEL-7B-MoT23.73#4619
11Z-Image23.67
12Qwen-Image23.53
13OmniGen221.8#4579
14FLUX.1-Krea-dev18.4

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