PureDocBench - Clean - Table TEDS: leaderboard
Metric: Table Tree-Edit-Distance-based Similarity (0-100) between the predicted and reference HTML tables, on the clean track (pages rendered straight from their HTML/CSS source) of 1,475 source pages in 10 domains and 66 subcategories, rendered from LLM-written HTML/CSS with source-linked annotations for text, formulas, tables and reading order; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 58 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 5 | 92.17 |
| 2 | GLM-5.3 Flash | 90.76 |
| 3 | Qwen 3.8 27B | 87.69 |
| 4 | Gemini 3.6 Flash | 86.76 |
| 5 | Qwen 3.5 122B A10B | 83.03 |
| 6 | GPT-5.6 Sol | 82.87 |
| 7 | Kimi K2.6 (Non-reasoning) | 80.3 |
| 8 | Qwen 3.5 9B | 79.39 |
| 9 | Qwen 3.5 4B | 78.02 |
| 10 | Qwen3.8-Flash-Next | 76.93 |
| 11 | Qwen 3.6 27B | 76.47 |
| 12 | Qwen 3.6 35B A3B | 76.25 |
| 13 | Gemini 3.1 Pro (Preview) | 75.08 |
| 14 | Qwen 3.5 27B | 72.51 |
| 15 | Qwen 3.5 2B | 70.7 |
Interactive version: theaggregate.ai/benchmark?slug=puredocbench-clean-table-teds · How It Works · Data refreshed daily, snapshot 2026-10-07.