KVBench (Detailed Caption) - Geography (Chinese): leaderboard
KVBench (2026) tests whether text-to-image models can draw correct senior high-school geography content, such as diagrams and scientific illustrations, from Chinese prompts given as a detailed caption that spells out the required visual content. 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 geography prompts in Chinese, each given as a detailed caption that spells out the required visual content: 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. 16 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Nano Banana Pro (Gemini 3 Pro Image) | 74.29 | |
| 2 | GPT-Image (KVBench checkpoint unspecified) | 43 | |
| 3 | Seedream 4.0 | 36.93 | |
| 4 | Qwen-Image | 23.11 | |
| 5 | Gemini 2.5 Flash Image | 21.77 | |
| 6 | FLUX.2-dev | 21.47 | |
| 7 | Z-Image | 18.59 | |
| 8 | FLUX 2 Max | 13.59 | |
| 9 | HiDream (KVBench checkpoint unspecified) | 7.84 | |
| 10 | Janus-Pro | 6.88 | |
| 11 | OmniGen2 | 5.64 | #4579 |
| 12 | SD3.5-Large (KVBench GGUF build, quantization unspecified) | 5.52 | |
| 13 | FLUX.1-Krea-dev | 5.33 | |
| 14 | FLUX.1-dev | 5.29 | |
| 15 | BAGEL-7B-MoT | 4.57 | #4619 |
Interactive version: theaggregate.ai/benchmark?slug=kvbench-detailed-caption-geography-chinese · How It Works · Data refreshed daily, snapshot 2026-11-21.