# Video Understanding on VideoMME (Score %) Source: [Video Understanding on VideoMME (Score %)]() Collection index: [llms.txt]() Data updated: 2026-07-01T07:40:15.198Z Evaluation of video understanding capabilities on the VideoMME dataset, reporting the primary 'Score (%)' metric. 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) | VideoMME | | Metric | Score (%) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 28 of 28 | | 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 - [Score (%)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Score (%) | Result date | Report paper | | --- | --- | --- | --- | | Gemini-1.5-Pro | 75 | Not recorded | 6a449ef1e0224bd4f4992e07 | | GPT-4o | 71.9 | Not recorded | 6a449ef1e0224bd4f4992e07 | | Gemini-1.5-Flash | 70.3 | Not recorded | 6a449ef1e0224bd4f4992e07 | | Vega | 69.4 | Not recorded | 6a449ef1e0224bd4f4992e07 | | Ouro-Spatial-8B | 66.5 | Not recorded | 6a2a4138d33dfb23e7839462 | | VideoLLaMA3 | 66.2 | Not recorded | 6a449ef1e0224bd4f4992e07 | | Qwen3-VL-8B | 65.5 | Not recorded | 6a2a4138d33dfb23e7839462 | | Qwen2.5-VL | 65.1 | Not recorded | 6a279e0e90966ade919bf69e | | Ouro-Spatial-4B | 63.9 | Not recorded | 6a2a4138d33dfb23e7839462 | | ProtoKV | 63.7 | Not recorded | 6a3e07c1206542251d7cb69a | | LLaVA-Video | 63.3 | Not recorded | 6a449ef1e0224bd4f4992e07 | | Qwen3-VL-4B | 62.9 | Not recorded | 6a2a4138d33dfb23e7839462 | | InfiniPot-V | 62.8 | Not recorded | 6a3e07c1206542251d7cb69a | | MOSS-Video-Preview | 62.48 | Not recorded | 6a279e0e90966ade919bf69e | | ProtoKV | 62.1 | Not recorded | 6a3e07c1206542251d7cb69a | | InfiniPot-V | 61.2 | Not recorded | 6a3e07c1206542251d7cb69a | | MOSS-Video-Preview | 59.81 | Not recorded | 6a279e0e90966ade919bf69e | | DyCoke | 59.7 | Not recorded | 6a3e07c1206542251d7cb69a | | LLaVA-OV | 58.2 | Not recorded | 6a449ef1e0224bd4f4992e07 | | Show-o2 | 57.4 | Not recorded | 6a449ef1e0224bd4f4992e07 | | GPT-4V | 57 | Not recorded | 6a449ef1e0224bd4f4992e07 | | LLaDA-V+Ours | 56.56 | Not recorded | 6a0d4054f984739ed6737b45 | | LLaDA-V | 56.07 | Not recorded | 6a0d4054f984739ed6737b45 | | LLaDA-V+XAtt | 55.63 | Not recorded | 6a0d4054f984739ed6737b45 | | DyCoke | 55.3 | Not recorded | 6a3e07c1206542251d7cb69a | | LongVA | 52.6 | Not recorded | 6a449ef1e0224bd4f4992e07 | | LLaDA-V+Flex | 48.22 | Not recorded | 6a0d4054f984739ed6737b45 | | VideoLLaMA2 | 47.9 | Not recorded | 6a449ef1e0224bd4f4992e07 | ### Gemini-1.5-Pro Score (%): 75. Result ID: 6a44c45e73d097d845be98ed. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"#Params.":"-"}. This record does not assert independent reproduction. ### GPT-4o Score (%): 71.9. Result ID: 6a44c45e73d097d845be98eb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"#Params.":"-"}. This record does not assert independent reproduction. ### Gemini-1.5-Flash Score (%): 70.3. Result ID: 6a44c45e73d097d845be98ec. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"#Params.":"-"}. This record does not assert independent reproduction. ### Vega Score (%): 69.4. Result ID: 6a44c45e73d097d845be98f4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"#Params.":"3B"}. This record does not assert independent reproduction. ### Ouro-Spatial-8B Score (%): 66.5. Result ID: 6a2a8d95fbc373592c286670. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model scale":"8B","Number of uniformly sampled frames":"32"}. This record does not assert independent reproduction. ### VideoLLaMA3 Score (%): 66.2. Result ID: 6a44c45e73d097d845be98f0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"#Params.":"7B"}. This record does not assert independent reproduction. ### Qwen3-VL-8B Score (%): 65.5. Result ID: 6a2a8d95fbc373592c28666f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model scale":"8B","Number of uniformly sampled frames":"32"}. This record does not assert independent reproduction. ### Qwen2.5-VL Score (%): 65.1. Result ID: 6a27b423fbc373592c23c9bf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### Ouro-Spatial-4B Score (%): 63.9. Result ID: 6a2a8d95fbc373592c28666e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model scale":"4B","Number of uniformly sampled frames":"32"}. This record does not assert independent reproduction. ### ProtoKV Score (%): 63.7. Result ID: 6a3e35c1fbc373592c6008b1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Budget \|M\|":"6K","Backbone":"Qwen2-VL-7B"}. This record does not assert independent reproduction. ### LLaVA-Video Score (%): 63.3. Result ID: 6a44c45e73d097d845be98f2. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"#Params.":"7B"}. This record does not assert independent reproduction. ### Qwen3-VL-4B Score (%): 62.9. Result ID: 6a2a8d95fbc373592c28666d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model scale":"4B","Number of uniformly sampled frames":"32"}. This record does not assert independent reproduction. ### InfiniPot-V Score (%): 62.8. Result ID: 6a3e35c1fbc373592c6008af. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Budget \|M\|":"6K","Backbone":"Qwen2-VL-7B"}. This record does not assert independent reproduction. ### MOSS-Video-Preview Score (%): 62.48. Result ID: 6a27b423fbc373592c23c9bd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"SFT Mode":"real-time"}. This record does not assert independent reproduction. ### ProtoKV Score (%): 62.1. Result ID: 6a3e35c1fbc373592c6008b0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Budget \|M\|":"3K","Backbone":"Qwen2-VL-7B"}. This record does not assert independent reproduction. ### InfiniPot-V Score (%): 61.2. Result ID: 6a3e35c1fbc373592c6008ae. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Budget \|M\|":"3K","Backbone":"Qwen2-VL-7B"}. This record does not assert independent reproduction. ### MOSS-Video-Preview Score (%): 59.81. Result ID: 6a27b423fbc373592c23c9be. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"SFT Mode":"offline"}. This record does not assert independent reproduction. ### DyCoke Score (%): 59.7. Result ID: 6a3e35c1fbc373592c6008ad. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Budget \|M\|":"6K","Backbone":"Qwen2-VL-7B"}. This record does not assert independent reproduction. ### LLaVA-OV Score (%): 58.2. Result ID: 6a44c45e73d097d845be98f1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"#Params.":"7B"}. This record does not assert independent reproduction. ### Show-o2 Score (%): 57.4. Result ID: 6a44c45e73d097d845be98f3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"#Params.":"7B"}. This record does not assert independent reproduction. ### GPT-4V Score (%): 57. Result ID: 6a44c45e73d097d845be98ea. [Reporting paper]() Recorded evaluation configuration: {"#Params.":"-"}. This record does not assert independent reproduction. ### LLaDA-V+Ours Score (%): 56.56. Result ID: 6a0d7a04fbc373592c05b86c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Attention":"Block Approximate Sparse Attention"}. This record does not assert independent reproduction. ### LLaDA-V Score (%): 56.07. Result ID: 6a0d7a04fbc373592c05b869. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Attention":"Full-Attention"}. This record does not assert independent reproduction. ### LLaDA-V+XAtt Score (%): 55.63. Result ID: 6a0d7a04fbc373592c05b86b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Attention":"XAtt"}. This record does not assert independent reproduction. ### DyCoke Score (%): 55.3. Result ID: 6a3e35c1fbc373592c6008ac. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Budget \|M\|":"3K","Backbone":"Qwen2-VL-7B"}. This record does not assert independent reproduction. ### LongVA Score (%): 52.6. Result ID: 6a44c45e73d097d845be98ee. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"#Params.":"7B"}. This record does not assert independent reproduction. ### LLaDA-V+Flex Score (%): 48.22. Result ID: 6a0d7a04fbc373592c05b86a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Attention":"Flex"}. This record does not assert independent reproduction. ### VideoLLaMA2 Score (%): 47.9. Result ID: 6a44c45e73d097d845be98ef. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"#Params.":"7B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.