# Video Understanding on MVBench, EgoSchema, NExT-QA, and VideoMME Aggregate Source: [Video Understanding on MVBench, EgoSchema, NExT-QA, and VideoMME Aggregate]() Collection index: [llms.txt]() Data updated: 2026-04-03T08:44:57.583Z Evaluation of general video understanding capabilities across the aggregated MVBench, EgoSchema, NExT-QA, and VideoMME datasets. 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 Understanding | | Dataset (including any stated split) | Aggregate (MVBench, EgoSchema, NExT-QA, VideoMME) | | Metric | Aggregate Score | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 19 of 19 | | 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 - [Aggregate Score](): Higher is better; unit: SCALAR. - [Aggregate Success Rate](): Higher is better; unit: PERCENT. ## Recorded results | Method | Aggregate Score | Result date | Report paper | | --- | --- | --- | --- | | LLaVA-OneVision | 56.4 | Not recorded | 69cf49b00636e5bf4afcdae9 | | HieraVid | 56.1 | Not recorded | 69cf49b00636e5bf4afcdae9 | | Framefusion | 55.6 | Not recorded | 69cf49b00636e5bf4afcdae9 | | HieraVid | 55.2 | Not recorded | 69cf49b00636e5bf4afcdae9 | | PruneVid | 54.9 | Not recorded | 69cf49b00636e5bf4afcdae9 | | Framefusion | 54.2 | Not recorded | 69cf49b00636e5bf4afcdae9 | | FastV | 53.9 | Not recorded | 69cf49b00636e5bf4afcdae9 | | PruMerge | 53.8 | Not recorded | 69cf49b00636e5bf4afcdae9 | | PruneVid | 53.6 | Not recorded | 69cf49b00636e5bf4afcdae9 | | HieraVid | 53.5 | Not recorded | 69cf49b00636e5bf4afcdae9 | | PruneVid | 52.3 | Not recorded | 69cf49b00636e5bf4afcdae9 | | Framefusion | 52 | Not recorded | 69cf49b00636e5bf4afcdae9 | | PruMerge | 51.6 | Not recorded | 69cf49b00636e5bf4afcdae9 | | FastV | 51.5 | Not recorded | 69cf49b00636e5bf4afcdae9 | | StreamingLLM | 51.1 | Not recorded | 69cf49b00636e5bf4afcdae9 | | PruMerge | 49.2 | Not recorded | 69cf49b00636e5bf4afcdae9 | | FastV | 49.1 | Not recorded | 69cf49b00636e5bf4afcdae9 | | StreamingLLM | 48.6 | Not recorded | 69cf49b00636e5bf4afcdae9 | | StreamingLLM | 46.9 | Not recorded | 69cf49b00636e5bf4afcdae9 | ### LLaVA-OneVision Aggregate Score: 56.4. Result ID: 69cf7deb47f34ce924b96ed8. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Relative token budget":"100%","Prefilling FLOPs (T)":"40.8","FLOPs Ratio":"100%"}. This record does not assert independent reproduction. ### HieraVid Aggregate Score: 56.1. Result ID: 69cf7deb47f34ce924b96ede. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"30%","Prefilling FLOPs (T)":"10.7","FLOPs Ratio":"26.3%"}. This record does not assert independent reproduction. ### Framefusion Aggregate Score: 55.6. Result ID: 69cf7deb47f34ce924b96edd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"30%","Prefilling FLOPs (T)":"10.6","FLOPs Ratio":"25.9%"}. This record does not assert independent reproduction. ### HieraVid Aggregate Score: 55.2. Result ID: 69cf7deb47f34ce924b96ee4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"20%","Prefilling FLOPs (T)":"7.5","FLOPs Ratio":"18.5%"}. This record does not assert independent reproduction. ### PruneVid Aggregate Score: 54.9. Result ID: 69cf7deb47f34ce924b96edc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"30%","Prefilling FLOPs (T)":"10.6","FLOPs Ratio":"25.9%"}. This record does not assert independent reproduction. ### Framefusion Aggregate Score: 54.2. Result ID: 69cf7deb47f34ce924b96ee3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"20%","Prefilling FLOPs (T)":"6.9","FLOPs Ratio":"16.9%"}. This record does not assert independent reproduction. ### FastV Aggregate Score: 53.9. Result ID: 69cf7deb47f34ce924b96eda. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"30%","Prefilling FLOPs (T)":"12.7","FLOPs Ratio":"31.2%"}. This record does not assert independent reproduction. ### PruMerge Aggregate Score: 53.8. Result ID: 69cf7deb47f34ce924b96edb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"30%","Prefilling FLOPs (T)":"10.6","FLOPs Ratio":"25.9%"}. This record does not assert independent reproduction. ### PruneVid Aggregate Score: 53.6. Result ID: 69cf7deb47f34ce924b96ee2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"20%","Prefilling FLOPs (T)":"6.9","FLOPs Ratio":"16.9%"}. This record does not assert independent reproduction. ### HieraVid Aggregate Score: 53.5. Result ID: 69cf7deb47f34ce924b96eea. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"10%","Prefilling FLOPs (T)":"3.8","FLOPs Ratio":"9.3%"}. This record does not assert independent reproduction. ### PruneVid Aggregate Score: 52.3. Result ID: 69cf7deb47f34ce924b96ee8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"10%","Prefilling FLOPs (T)":"4.3","FLOPs Ratio":"10.6%"}. This record does not assert independent reproduction. ### Framefusion Aggregate Score: 52. Result ID: 69cf7deb47f34ce924b96ee9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"10%","Prefilling FLOPs (T)":"3.4","FLOPs Ratio":"8.4%"}. This record does not assert independent reproduction. ### PruMerge Aggregate Score: 51.6. Result ID: 69cf7deb47f34ce924b96ee1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"20%","Prefilling FLOPs (T)":"6.9","FLOPs Ratio":"16.9%"}. This record does not assert independent reproduction. ### FastV Aggregate Score: 51.5. Result ID: 69cf7deb47f34ce924b96ee0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"20%","Prefilling FLOPs (T)":"9.3","FLOPs Ratio":"22.8%"}. This record does not assert independent reproduction. ### StreamingLLM Aggregate Score: 51.1. Result ID: 69cf7deb47f34ce924b96ed9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"30%","Prefilling FLOPs (T)":"30.6","FLOPs Ratio":"75%"}. This record does not assert independent reproduction. ### PruMerge Aggregate Score: 49.2. Result ID: 69cf7deb47f34ce924b96ee7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"10%","Prefilling FLOPs (T)":"3.4","FLOPs Ratio":"8.3%"}. This record does not assert independent reproduction. ### FastV Aggregate Score: 49.1. Result ID: 69cf7deb47f34ce924b96ee6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"10%","Prefilling FLOPs (T)":"6.037","FLOPs Ratio":"14.8%"}. This record does not assert independent reproduction. ### StreamingLLM Aggregate Score: 48.6. Result ID: 69cf7deb47f34ce924b96edf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"20%","Prefilling FLOPs (T)":"29.5","FLOPs Ratio":"72.3%"}. This record does not assert independent reproduction. ### StreamingLLM Aggregate Score: 46.9. Result ID: 69cf7deb47f34ce924b96ee5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Relative token budget":"10%","Prefilling FLOPs (T)":"26.9","FLOPs Ratio":"65.9%"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.