# Visual Question Answering on DocVQA (held-out)

Source: [Visual Question Answering on DocVQA (held-out)](<https://www.sota2.com/research/sota/visual-question-answering-on-docvqa-held-out>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:22:33.801Z

Evaluation of model performance on the held-out split of the DocVQA dataset for Visual Question Answering, measured using ANLS.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Visual Question Answering |
| Dataset (including any stated split) | DocVQA (held-out) |
| Metric | ANLS |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [ANLS](<https://www.sota2.com/research/sota/visual-question-answering-on-docvqa-held-out.md?metric=ANLS>): Higher is better; unit: PERCENT.

## Recorded results

| Method | ANLS | Result date | Report paper |
| --- | --- | --- | --- |
| GPT-4+OCR | 82.8 | Not recorded | 68fa71b61e209131cea52b22 |
| DocLLM-7B | 63.4 | Not recorded | 68fa71b61e209131cea52b22 |
| DocLLM-1B | 53.5 | Not recorded | 68fa71b61e209131cea52b22 |
| Llama2+OCR | 47.4 | Not recorded | 68fa71b61e209131cea52b22 |

### GPT-4+OCR

ANLS: 82.8. Result ID: 6991f78e3a6c3148490bc52d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71b61e209131cea52b22.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa727a1e209131cea56a1e.md>)

[Implementation](<https://github.com/openai/evals>)

Recorded evaluation configuration: {"Size":"~1T","Setting":"ZS"}. This record does not assert independent reproduction.

### DocLLM-7B

ANLS: 63.4. Result ID: 6991f78e3a6c3148490bc530.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71b61e209131cea52b22.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71b61e209131cea52b22.md>)

Recorded evaluation configuration: {"Size":"7B","Setting":"STDD"}. This record does not assert independent reproduction.

### DocLLM-1B

ANLS: 53.5. Result ID: 6991f78e3a6c3148490bc52f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71b61e209131cea52b22.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71b61e209131cea52b22.md>)

Recorded evaluation configuration: {"Size":"1B","Setting":"STDD"}. This record does not assert independent reproduction.

### Llama2+OCR

ANLS: 47.4. Result ID: 6991f78e3a6c3148490bc52e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71b61e209131cea52b22.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa701f1e209131cea492fe.md>)

[Implementation](<https://github.com/facebookresearch/llama>)

Recorded evaluation configuration: {"Size":"7B","Setting":"ZS"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
