DocVQAtest
Progress Over Time
Interactive timeline showing model performance evolution on DocVQAtest
DocVQAtest Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  Alibaba Cloud / Qwen Team | 236B | — | — | ||
| 2 | Alibaba Cloud / Qwen Team | 33B | — | — | ||
| 3 | Alibaba Cloud / Qwen Team | 73B | — | — | ||
| 3 | Alibaba Cloud / Qwen Team | 236B | — | — | ||
| 5 | Alibaba Cloud / Qwen Team | 33B | — | — | ||
| 5 | Alibaba Cloud / Qwen Team | 9B | — | — | ||
| 7 | Alibaba Cloud / Qwen Team | 4B | 262K | $0.10 / $0.60 | ||
| 7 | Alibaba Cloud / Qwen Team | 9B | — | — | ||
| 9 | Alibaba Cloud / Qwen Team | 31B | — | — | ||
| 9 | Alibaba Cloud / Qwen Team | 31B | — | — | ||
| 11 | Alibaba Cloud / Qwen Team | 4B | 262K | $0.10 / $1.00 |
What is DocVQAtest?
DocVQA is a Visual Question Answering benchmark on document images containing 50,000 questions defined on 12,000+ document images. The benchmark focuses on understanding document structure and content to answer questions about various document types including letters, memos, notes, and reports from the UCSF Industry Documents Library.
DocVQAtest is a multimodal benchmark evaluating models on multimodal and vision tasks. LLM Stats tracks 11 models on this benchmark, scored on a 0–1 scale. The current average is 1.0, with the leader at 1.0.
Compare leaders on the best AI for multimodal and best AI for vision leaderboards.
Current leaders
Qwen3 VL 235B A22B Instruct from Alibaba Cloud / Qwen Team currently leads the DocVQAtest leaderboard with a score of 0.971 across 11 evaluated AI models.
Source paper
- Title
- DocVQA: A Dataset for VQA on Document Images
- Authors
- Minesh Mathew, Dimosthenis Karatzas, C. V. Jawahar
- Published
- arXiv
- 2007.00398
Abstract
We present a new dataset for Visual Question Answering (VQA) on document images called DocVQA. The dataset consists of 50,000 questions defined on 12,000+ document images. Detailed analysis of the dataset in comparison with similar datasets for VQA and reading comprehension is presented. We report several baseline results by adopting existing VQA and reading comprehension models. Although the existing models perform reasonably well on certain types of questions, there is large performance gap compared to human performance (94.36% accuracy). The models need to improve specifically on questions where understanding structure of the document is crucial. The dataset, code and leaderboard are available at docvqa.org
FAQ
Common questions about the DocVQAtest benchmark and leaderboard.