CiteVQA - Multi-Doc (N-Gold): leaderboard
Metric: Strict attributed accuracy (%): share of questions with answer score at least 4 and either relevance at least 4 or evidence recall at least 0.6, on the multi-document questions with several gold documents (22.3 percent of the set), the model answers from the document pages and cites element-level bounding boxes, unified prompt, temperature 1.0, at most 4,096 output tokens, Qwen3-VL-235B-A22B judge; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) (High) | 71.6 |
| 2 | Gemini 3 Flash (Preview) (High) | 60.5 |
| 3 | GPT-5.4 (xHigh) | 55.1 |
| 4 | Seed 2.0 Pro | 36.2 |
| 5 | GPT-5.2 (xHigh) | 30.5 |
| 6 | Gemma 4 31B | 17.8 |
| 7 | Qwen 3 VL 235B A22B Instruct | 17.8 |
| 8 | Qwen 3.5 397B A17B | 15.2 |
| 9 | Kimi K2.5 | 14.2 |
| 10 | Qwen 3 VL 32B Instruct | 14 |
| 11 | GLM-5V Turbo | 13 |
| 12 | Qwen 3.5 27B | 11.6 |
| 13 | Qwen 3.5 9B | 10.7 |
| 14 | Qwen 3.6 Plus | 9.8 |
| 15 | Qwen 3.5 122B A10B | 9.2 |
Interactive version: theaggregate.ai/benchmark?slug=citevqa-multi-doc-n-gold · How It Works · Data refreshed daily, snapshot 2026-10-07.