# Document Visual Question Answering on DocVQA (DocVQA Metric)

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

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

Data updated: 2026-07-03T07:29:20.577Z

Evaluation of model performance on the DocVQA dataset for Document Visual Question Answering, reporting the metric specifically labeled 'DocVQA'.

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 | Document Visual Question Answering |
| Dataset (including any stated split) | DocVQA |
| Metric | DocVQA Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 18 of 18 |
| 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

- [DocVQA Score](<https://www.sota2.com/research/sota/document-visual-question-answering-on-docvqa-docvqa-metric.md?metric=DocVQA%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | DocVQA Score | Result date | Report paper |
| --- | --- | --- | --- |
| GPTQ | 94.01 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| Full Precision | 94 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| Full Precision | 93.4 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| Full Precision | 92.47 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| GPTQ | 90.25 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| GPTQ | 89.82 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| SAB-LVLM | 87.19 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| ARB-LLM | 86.85 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| ARB-LLM | 85.15 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| SAB-LVLM | 84.34 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| PB-LLM | 54.02 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| SAB-LVLM | 48.03 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| BiLLM | 47.09 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| BiLLM | 42.9 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| ARB-LLM | 40.67 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| BiLLM | 0.93 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| PB-LLM | 0.03 | Not recorded | 6a47420e143ebdcacacc5cf9 |
| PB-LLM | 0 | Not recorded | 6a47420e143ebdcacacc5cf9 |

### GPTQ

DocVQA Score: 94.01. Result ID: 6a4764cf60b88bb33992d4ba.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ecc1e209131cea4235d.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-14B-Instruct","Weight bits":"3 bit","Memory":"5.38 GB"}. This record does not assert independent reproduction.

### Full Precision

DocVQA Score: 94. Result ID: 6a4764cf60b88bb33992d4bf.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-38B-Instruct","Weight bits":"16 bit","Memory":"71.51 GB"}. This record does not assert independent reproduction.

### Full Precision

DocVQA Score: 93.4. Result ID: 6a4764cf60b88bb33992d4b9.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-14B-Instruct","Weight bits":"16 bit","Memory":"28.16 GB"}. This record does not assert independent reproduction.

### Full Precision

DocVQA Score: 92.47. Result ID: 6a4764cf60b88bb33992d4b3.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-8B-Instruct","Weight bits":"16 bit","Memory":"15.89 GB"}. This record does not assert independent reproduction.

### GPTQ

DocVQA Score: 90.25. Result ID: 6a4764cf60b88bb33992d4c0.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ecc1e209131cea4235d.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-38B-Instruct","Weight bits":"3 bit","Memory":"12.71 GB"}. This record does not assert independent reproduction.

### GPTQ

DocVQA Score: 89.82. Result ID: 6a4764cf60b88bb33992d4b4.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ecc1e209131cea4235d.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-8B-Instruct","Weight bits":"3 bit","Memory":"2.83 GB"}. This record does not assert independent reproduction.

### SAB-LVLM

DocVQA Score: 87.19. Result ID: 6a4764cf60b88bb33992d4c4.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-38B-Instruct","Weight bits":"1.06 bit","Memory":"10.27 GB"}. This record does not assert independent reproduction.

### ARB-LLM

DocVQA Score: 86.85. Result ID: 6a4764cf60b88bb33992d4c3.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa751e1e209131cea65e31.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-38B-Instruct","Weight bits":"1.08 bit","Memory":"10.27 GB"}. This record does not assert independent reproduction.

### ARB-LLM

DocVQA Score: 85.15. Result ID: 6a4764cf60b88bb33992d4bd.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa751e1e209131cea65e31.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-14B-Instruct","Weight bits":"1.07 bit","Memory":"4.33 GB"}. This record does not assert independent reproduction.

### SAB-LVLM

DocVQA Score: 84.34. Result ID: 6a4764cf60b88bb33992d4be.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-14B-Instruct","Weight bits":"1.07 bit","Memory":"4.33 GB"}. This record does not assert independent reproduction.

### PB-LLM

DocVQA Score: 54.02. Result ID: 6a4764cf60b88bb33992d4c1.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa711d1e209131cea4f306.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-38B-Instruct","Weight bits":"1.70 bit","Memory":"10.24 GB"}. This record does not assert independent reproduction.

### SAB-LVLM

DocVQA Score: 48.03. Result ID: 6a4764cf60b88bb33992d4b8.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-8B-Instruct","Weight bits":"1.07 bit","Memory":"2.28 GB"}. This record does not assert independent reproduction.

### BiLLM

DocVQA Score: 47.09. Result ID: 6a4764cf60b88bb33992d4bc.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73a81e209131cea5db8c.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-14B-Instruct","Weight bits":"1.07 bit","Memory":"4.53 GB"}. This record does not assert independent reproduction.

### BiLLM

DocVQA Score: 42.9. Result ID: 6a4764cf60b88bb33992d4c2.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73a81e209131cea5db8c.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-38B-Instruct","Weight bits":"1.07 bit","Memory":"10.72 GB"}. This record does not assert independent reproduction.

### ARB-LLM

DocVQA Score: 40.67. Result ID: 6a4764cf60b88bb33992d4b7.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa751e1e209131cea65e31.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-8B-Instruct","Weight bits":"1.07 bit","Memory":"2.28 GB"}. This record does not assert independent reproduction.

### BiLLM

DocVQA Score: 0.93. Result ID: 6a4764cf60b88bb33992d4b6.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73a81e209131cea5db8c.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-8B-Instruct","Weight bits":"1.08 bit","Memory":"2.39 GB"}. This record does not assert independent reproduction.

### PB-LLM

DocVQA Score: 0.03. Result ID: 6a4764cf60b88bb33992d4b5.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa711d1e209131cea4f306.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-8B-Instruct","Weight bits":"1.70 bit","Memory":"2.28 GB"}. This record does not assert independent reproduction.

### PB-LLM

DocVQA Score: 0. Result ID: 6a4764cf60b88bb33992d4bb.

[Reporting paper](<https://www.sota2.com/research/paper/6a47420e143ebdcacacc5cf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa711d1e209131cea4f306.md>)

Recorded evaluation configuration: {"Model":"InternVL3.5-14B-Instruct","Weight bits":"1.70 bit","Memory":"4.34 GB"}. This record does not assert independent reproduction.

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