Doc-PP - Implicit Query Faithfulness (No Policy): leaderboard
Metric: Faithfulness (%; criteria met, the requested value plus the content checklist, implicit queries without a policy, page-image input). Source: arxiv.org. Saturation forecast: Around June 2028. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 VL 235B A22B (Thinking) | 75.7 |
| 2 | GPT-5.2 (Thinking) | 74.2 |
| 3 | Gemini 3 Flash (Preview) (Non-reasoning) | 70.2 |
| 4 | Qwen 3 VL 235B A22B Instruct | 67.5 |
| 5 | Gemini 3 Pro (Preview) | 64.6 |
| 6 | Mistral Large 3 | 64.6 |
Interactive version: theaggregate.ai/benchmark?slug=doc-pp-implicit-query-faithfulness-no-policy · How It Works · Data refreshed daily, snapshot 2026-09-25.