DTBench - Direct Extraction Recall: leaderboard
Metric: Recall (%) on DTBench's 4,293 directly extractable ground-truth cells (values stated verbatim in the synthesized document), exact match after normalization and row alignment; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V3.2 | 96.37 | #198 |
| 2 | Gemini 3 Flash | 95.93 | #93 |
| 3 | GPT-5 Mini | 92.93 | #176 |
| 4 | GPT-5 | 92.48 | #91 |
| 5 | Qwen 3 32B | 73.84 | #424 |
| 6 | Qwen 3 4B | 35.47 | #823 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=dtbench-direct-extraction-recall · How It Works · Data refreshed daily, snapshot 2026-10-11.