multimodal benchmark
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.
Updated Aug 7, 2026
Published models11
Registry coverage11
MetricScore
EvidenceB
DocVQAtest leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AC | 97.1% | 100.0% | 11 | C | |
| 2 | AC | 96.9% | 90.0% | 11 | C | |
| 3 | AC | 96.5% | 80.0% | 11 | C | |
| 4 | AC | 96.5% | 70.0% | 11 | C | |
| 5 | AC | 96.1% | 60.0% | 11 | C | |
| 6 | AC | 96.1% | 50.0% | 11 | C | |
| 7 | AC | 95.3% | 40.0% | 11 | C | |
| 8 | AC | 95.3% | 30.0% | 11 | C | |
| 9 | AC | 95.0% | 20.0% | 11 | C | |
| 10 | AC | 95.0% | 10.0% | 11 | C | |
| 11 | AC | 94.2% | 0.0% | 11 | C |
Score distribution
Top published rows on the benchmark's original scale.
DocVQAtest highlights
The top published results on this benchmark's own scale.
What is DocVQAtest?
Definition and scoring fields from the benchmark registry.
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.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- DocVQAtest
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- docvqatest|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about DocVQAtest.
Which model scores highest on DocVQAtest?
Qwen3 VL 235B A22B Instruct is currently ranked first with 97.1%.
What does DocVQAtest measure?
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.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
11 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.