# Visual Question Answering on VQAv2 Reliable Challenge (test-dev)

Source: [Visual Question Answering on VQAv2 Reliable Challenge (test-dev)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqav2-reliable-challenge-test-dev>)

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

Data updated: 2026-02-26T05:31:52.040Z

Evaluation of model performance on the test-development split of the VQAv2 Reliable Challenge dataset for visual question answering, reporting Accuracy, Phi100, and Coverage metrics.

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) | VQAv2 Reliable Challenge (test-dev) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 1 of 1 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/visual-question-answering-on-vqav2-reliable-challenge-test-dev.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Phi100](<https://www.sota2.com/research/sota/visual-question-answering-on-vqav2-reliable-challenge-test-dev.md?metric=Phi100>): Higher is better; unit: PERCENT.

- [AUC (100x)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqav2-reliable-challenge-test-dev.md?metric=AUC%20(100x)>): Higher is better; unit: PERCENT.

- [Coverage @ 0.5%](<https://www.sota2.com/research/sota/visual-question-answering-on-vqav2-reliable-challenge-test-dev.md?metric=Coverage%20%40%200.5%25>): Higher is better; unit: PERCENT.

- [Coverage @ 1%](<https://www.sota2.com/research/sota/visual-question-answering-on-vqav2-reliable-challenge-test-dev.md?metric=Coverage%20%40%201%25>): Higher is better; unit: PERCENT.

- [Coverage @ 2%](<https://www.sota2.com/research/sota/visual-question-answering-on-vqav2-reliable-challenge-test-dev.md?metric=Coverage%20%40%202%25>): Higher is better; unit: PERCENT.

- [Coverage @ 5%](<https://www.sota2.com/research/sota/visual-question-answering-on-vqav2-reliable-challenge-test-dev.md?metric=Coverage%20%40%205%25>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| DAVR | 83.99 | Not recorded | 69438a89423ea42265729455 |

### DAVR

Accuracy: 83.99. Result ID: 69983f865698665ec49bda1f.

[Reporting paper](<https://www.sota2.com/research/paper/69438a89423ea42265729455.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69438a89423ea42265729455.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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