WildTableBench: leaderboard
WildTableBench (2026) asks multimodal models 928 human-verified free-form questions about 402 table images collected from the web, in 17 subtypes and five categories. The score is overall accuracy in percent, judged against the reference answer by GPT-5.2; higher is better.
Metric: Overall accuracy (%) on WildTableBench, 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 February 2027. 21 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Gemini 3 Pro | 67.9 | #145 | |
| 2 | Kimi K2.5 | Thinking | 49.9 | #266 |
| 3 | Gemini 3 Flash | 49.4 | #175 | |
| 4 | Seed 2.0 Pro | High | 47.8 | #222 |
| 5 | GPT-5.2 | High | 46.6 | #185 |
| 6 | Claude Opus 4.6 | High | 45.7 | #113 |
| 7 | Claude Sonnet 4.6 | High | 35 | #173 |
| 8 | Qwen 3 VL 235B A22B (Thinking) | 34.7 | #459 | |
| 9 | Qwen 3 VL 32B (Thinking) | 28.2 | #573 | |
| 10 | GPT-5 Mini | High | 25.3 | #364 |
| 11 | Qwen 3 VL 235B A22B Instruct | 25.2 | #532 | |
| 12 | Qwen 3 VL 32B Instruct | 24.5 | #606 | |
| 13 | GLM-4.6V | 24.4 | #710 | |
| 14 | O3 | High | 14.8 | #237 |
| 15 | Qwen 3 VL 8B (Thinking) | 14.7 | #910 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=wildtablebench · How It Works · Data refreshed daily, snapshot 2026-11-21.