# Video Question Answering on MLVU (test) Source: [Video Question Answering on MLVU (test)]() Collection index: [llms.txt]() Data updated: 2026-04-08T07:47:16.283Z Evaluation of model performance on the test split of the MLVU dataset for Video Question Answering, measured using Accuracy. 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 | Video Question Answering | | Dataset (including any stated split) | MLVU (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | InternVL 2.5 | 75.7 | Not recorded | 69d5e1237d0172140ec75b34 | | Video-RAG | 73.8 | Not recorded | 69d5e1237d0172140ec75b34 | | Qwen2.5-VL | 73.1 | Not recorded | 69d5e1237d0172140ec75b34 | | LLaVA-Video | 72.9 | Not recorded | 69d5e1237d0172140ec75b34 | | Vgent | 72.1 | Not recorded | 69d5e1237d0172140ec75b34 | | SVAgent | 65.6 | Not recorded | 69d5e1237d0172140ec75b34 | | SVAgent | 64.9 | Not recorded | 69d5e1237d0172140ec75b34 | | GPT-4o | 64.6 | Not recorded | 69d5e1237d0172140ec75b34 | | VideoMind | 64.4 | Not recorded | 69d5e1237d0172140ec75b34 | | Gemini 1.5 Pro | 64 | Not recorded | 69d5e1237d0172140ec75b34 | | SVAgent | 62.7 | Not recorded | 69d5e1237d0172140ec75b34 | | SVAgent | 61.2 | Not recorded | 69d5e1237d0172140ec75b34 | | Qwen2.5-VL | 54.7 | Not recorded | 69d5e1237d0172140ec75b34 | | Qwen3-VL | 54.5 | Not recorded | 69d5e1237d0172140ec75b34 | | Qwen2.5-VL | 53.6 | Not recorded | 69d5e1237d0172140ec75b34 | | Qwen3-VL | 53.4 | Not recorded | 69d5e1237d0172140ec75b34 | ### InternVL 2.5 Accuracy: 75.7. Result ID: 69d607fb47f34ce924be50b2. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Size":"78B","# Frames":"64"}. This record does not assert independent reproduction. ### Video-RAG Accuracy: 73.8. Result ID: 69d607fb47f34ce924be50b5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Size":"72B","# Frames":"64"}. This record does not assert independent reproduction. ### Qwen2.5-VL Accuracy: 73.1. Result ID: 69d607fb47f34ce924be50b1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Size":"72B","# Frames":"32 / 768∗"}. This record does not assert independent reproduction. ### LLaVA-Video Accuracy: 72.9. Result ID: 69d607fb47f34ce924be50b0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Size":"72B","# Frames":"32 / 64∗"}. This record does not assert independent reproduction. ### Vgent Accuracy: 72.1. Result ID: 69d607fb47f34ce924be50b4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Size":"7B","# Frames":"320"}. This record does not assert independent reproduction. ### SVAgent Accuracy: 65.6. Result ID: 69d607fb47f34ce924be50bd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL","Size":"8B","# Frames":"8"}. This record does not assert independent reproduction. ### SVAgent Accuracy: 64.9. Result ID: 69d607fb47f34ce924be50b9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-VL","Size":"4B","# Frames":"8"}. This record does not assert independent reproduction. ### GPT-4o Accuracy: 64.6. Result ID: 69d607fb47f34ce924be50af. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Size":"-","# Frames":"1fps"}. This record does not assert independent reproduction. ### VideoMind Accuracy: 64.4. Result ID: 69d607fb47f34ce924be50b3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Size":"7B","# Frames":"32"}. This record does not assert independent reproduction. ### Gemini 1.5 Pro Accuracy: 64. Result ID: 69d607fb47f34ce924be50ae. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Size":"-","# Frames":"1fps"}. This record does not assert independent reproduction. ### SVAgent Accuracy: 62.7. Result ID: 69d607fb47f34ce924be50bb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Size":"7B","# Frames":"8"}. This record does not assert independent reproduction. ### SVAgent Accuracy: 61.2. Result ID: 69d607fb47f34ce924be50b7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL","Size":"3B","# Frames":"8"}. This record does not assert independent reproduction. ### Qwen2.5-VL Accuracy: 54.7. Result ID: 69d607fb47f34ce924be50ba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Size":"7B","# Frames":"8"}. This record does not assert independent reproduction. ### Qwen3-VL Accuracy: 54.5. Result ID: 69d607fb47f34ce924be50bc. [Reporting paper]() Recorded evaluation configuration: {"Size":"8B","# Frames":"8"}. This record does not assert independent reproduction. ### Qwen2.5-VL Accuracy: 53.6. Result ID: 69d607fb47f34ce924be50b6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Size":"3B","# Frames":"8"}. This record does not assert independent reproduction. ### Qwen3-VL Accuracy: 53.4. Result ID: 69d607fb47f34ce924be50b8. [Reporting paper]() Recorded evaluation configuration: {"Size":"4B","# Frames":"8"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.