DocScope - Page Localization: leaderboard
Metric: Page localization F1 (%): micro-averaged F1 of the predicted evidence pages against the gold pages by exact page match over the answerable questions of the 730-question DocScope test split (human-annotated questions on long, visually rich PDF documents; the model receives the whole document and returns evidence pages, evidence regions, supporting facts and an answer), each stage scored only on the gold pages the model retrieved; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 87.6 |
| 2 | Claude Opus 4.7 | 86 |
| 3 | Claude Sonnet 4.6 | 84 |
| 4 | Gemini 3.1 Flash Lite | 83.8 |
| 5 | Qwen 3.5 27B | 82.1 |
| 6 | Qwen 3.6 Plus | 78.9 |
| 7 | GPT-5.4 | 78.4 |
| 8 | Gemma 4 31B | 76.9 |
| 9 | Qwen 3.5 397B A17B | 74.1 |
| 10 | Qwen 3 VL 235B A22B | 72.1 |
| 11 | Ministral 3 14B | 58.4 |
| 12 | Gemma 4 26B A4B | 56.4 |
| 13 | Qwen 3.5 122B A10B | 55 |
| 14 | Ministral 3 8B | 53 |
Interactive version: theaggregate.ai/benchmark?slug=docscope-page-localization · How It Works · Data refreshed daily, snapshot 2026-10-07.