WildTableBench - Hypothetical Condition: leaderboard
Metric: Accuracy (%) on the Hypothetical Condition subtype (C4-H), 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 January 2027. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro | 82.1 |
| 2 | GPT-5.2 (High) | 60.7 |
| 3 | Kimi K2.5 (Thinking) | 53.6 |
| 4 | Claude Sonnet 4.6 (High) | 53.6 |
| 5 | Gemini 3 Flash | 50 |
| 6 | Qwen 3 VL 235B A22B (Thinking) | 50 |
| 7 | Claude Opus 4.6 (High) | 50 |
| 8 | Seed 2.0 Pro (High) | 42.9 |
| 9 | GPT-5 Mini (High) | 32.1 |
| 10 | Qwen 3 VL 8B (Thinking) | 25 |
| 11 | Qwen 3 VL 32B Instruct | 21.4 |
| 12 | GLM-4.6V | 17.9 |
| 13 | Qwen 3 VL 235B A22B Instruct | 14.3 |
| 14 | Qwen 3 VL 32B (Thinking) | 10.7 |
| 15 | O3 (High) | 10.7 |
Interactive version: theaggregate.ai/benchmark?slug=wildtablebench-hypothetical-condition · How It Works · Data refreshed daily, snapshot 2026-10-07.