WildTableBench - Color-based Counting: leaderboard
A WildTableBench (2026) subtype on counting the cells or records that carry a target color, asked about real-world table images from the web. Accuracy in percent, free-form answers judged against the reference by GPT-5.2; higher is better.
Metric: Accuracy (%) on the Color-based Counting subtype (C5-C), 928 human-verified questions over 402 real-world table images, free-form answers judged by GPT-5.2 against the reference answer, high reasoning effort where configurable; higher is better. Source: arxiv.org. Saturation forecast: Around March 2027. 21 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Gemini 3 Pro | 62.5 | #146 | |
| 2 | Gemini 3 Flash | 37.5 | #180 | |
| 3 | Seed 2.0 Pro | High | 37.5 | #226 |
| 4 | GPT-5.2 | High | 30 | #190 |
| 5 | Kimi K2.5 | Thinking | 27.5 | #272 |
| 6 | Qwen 3 VL 32B (Thinking) | 25 | #577 | |
| 7 | Claude Opus 4.6 | High | 25 | #114 |
| 8 | Qwen 3 VL 235B A22B (Thinking) | 22.5 | #462 | |
| 9 | Qwen 3 VL 32B Instruct | 20 | #607 | |
| 10 | GLM-4.6V | 17.5 | #721 | |
| 11 | Claude Sonnet 4.6 | High | 17.5 | #179 |
| 12 | Qwen 3 VL 235B A22B Instruct | 15 | #541 | |
| 13 | GPT-5 Mini | High | 15 | #373 |
| 14 | Qwen 3 VL 4B (Thinking) | 12.5 | #1354 | |
| 15 | Qwen 3 VL 8B (Thinking) | 10 | #924 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=wildtablebench-color-based-counting · How It Works · Data refreshed daily, snapshot 2026-11-22.