# Multimodal Question Answering on Encyclopedic-VQA 90 (test)

Source: [Multimodal Question Answering on Encyclopedic-VQA 90 (test)](<https://www.sota2.com/research/sota/multimodal-question-answering-on-encyclopedic-vqa-90-test>)

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

Data updated: 2026-05-29T09:23:19.525Z

Evaluation of model performance on the test split of the Encyclopedic-VQA 90 dataset for Multimodal Question Answering, reporting accuracy, coherence, RP, and latency.

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 | Multimodal Question Answering |
| Dataset (including any stated split) | Encyclopedic-VQA 90 (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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/multimodal-question-answering-on-encyclopedic-vqa-90-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Coherence](<https://www.sota2.com/research/sota/multimodal-question-answering-on-encyclopedic-vqa-90-test.md?metric=Coherence>): Higher is better; unit: PERCENT.

- [RP (%)](<https://www.sota2.com/research/sota/multimodal-question-answering-on-encyclopedic-vqa-90-test.md?metric=RP%20(%25)>): Higher is better; unit: PERCENT.

- [Latency (s)](<https://www.sota2.com/research/sota/multimodal-question-answering-on-encyclopedic-vqa-90-test.md?metric=Latency%20(s)>): Lower is better; unit: TIME.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| CogniVerse | 84.3 | Not recorded | 6a191d84dbd7150506f82794 |
| MMCoQA | 78.5 | Not recorded | 6a191d84dbd7150506f82794 |
| GraphRAG | 76.8 | Not recorded | 6a191d84dbd7150506f82794 |
| MuRAG | 74.2 | Not recorded | 6a191d84dbd7150506f82794 |
| BLIP-2 | 68.7 | Not recorded | 6a191d84dbd7150506f82794 |
| CLIP-ViT-L | 62.4 | Not recorded | 6a191d84dbd7150506f82794 |

### CogniVerse

Accuracy: 84.3. Result ID: 6a19597efbc373592c13f586.

[Reporting paper](<https://www.sota2.com/research/paper/6a191d84dbd7150506f82794.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a191d84dbd7150506f82794.md>)

Recorded evaluation configuration: {"Backbone":"LLaVA-13B","retrieval":"hyperbolic embedding space"}. This record does not assert independent reproduction.

### MMCoQA

Accuracy: 78.5. Result ID: 6a19597efbc373592c13f584.

[Reporting paper](<https://www.sota2.com/research/paper/6a191d84dbd7150506f82794.md>)

Recorded evaluation configuration: {"retrieval":"sequence-level objectives"}. This record does not assert independent reproduction.

### GraphRAG

Accuracy: 76.8. Result ID: 6a19597efbc373592c13f585.

[Reporting paper](<https://www.sota2.com/research/paper/6a191d84dbd7150506f82794.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa770d1e209131cea70c5a.md>)

Recorded evaluation configuration: {"retrieval":"static knowledge graphs"}. This record does not assert independent reproduction.

### MuRAG

Accuracy: 74.2. Result ID: 6a19597efbc373592c13f583.

[Reporting paper](<https://www.sota2.com/research/paper/6a191d84dbd7150506f82794.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70ec1e209131cea4e0cd.md>)

[Implementation](<https://open-vision-language.github.io/infoseek/>)

Recorded evaluation configuration: {"retrieval":"Euclidean embeddings"}. This record does not assert independent reproduction.

### BLIP-2

Accuracy: 68.7. Result ID: 6a19597efbc373592c13f582.

[Reporting paper](<https://www.sota2.com/research/paper/6a191d84dbd7150506f82794.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fb71e209131cea46fe6.md>)

[Implementation](<https://github.com/salesforce/LAVIS/tree/main/projects/blip2>)

Recorded evaluation configuration: {"retrieval":"none"}. This record does not assert independent reproduction.

### CLIP-ViT-L

Accuracy: 62.4. Result ID: 6a19597efbc373592c13f581.

[Reporting paper](<https://www.sota2.com/research/paper/6a191d84dbd7150506f82794.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"retrieval":"none","fine-tuned":"true"}. This record does not assert independent reproduction.

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