# Long-horizon video understanding on MM-Lifelong (train@month) Source: [Long-horizon video understanding on MM-Lifelong (train@month)]() Collection index: [llms.txt]() Data updated: 2026-05-19T07:32:32.367Z Evaluation of long-horizon video understanding capabilities on the train@month split of the MM-Lifelong dataset. 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 | Long-horizon video understanding | | Dataset (including any stated split) | MM-Lifelong (train@month) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 11 of 11 | | 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. - [Retrieval Rate@300](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Human | 82.5 | Not recorded | 6a0bee61dbc836b87d82c1cc | | ReMA | 17.62 | Not recorded | 6a0bee61dbc836b87d82c1cc | | VAM | 17.11 | Not recorded | 6a0bee61dbc836b87d82c1cc | | ReMA | 14.23 | Not recorded | 6a0bee61dbc836b87d82c1cc | | GPT-5 | 10.15 | Not recorded | 6a0bee61dbc836b87d82c1cc | | Qwen3-VL-235B-A22B | 9.09 | Not recorded | 6a0bee61dbc836b87d82c1cc | | Qwen3-VL-30B-A3B | 8.33 | Not recorded | 6a0bee61dbc836b87d82c1cc | | Nemotron-v2-12B | 7.52 | Not recorded | 6a0bee61dbc836b87d82c1cc | | Video-XL-2-8B | 6.02 | Not recorded | 6a0bee61dbc836b87d82c1cc | | LongVT-7B | 5.83 | Not recorded | 6a0bee61dbc836b87d82c1cc | | Eagle-2.5-8B | 3.76 | Not recorded | 6a0bee61dbc836b87d82c1cc | ### Human Accuracy: 82.5. Result ID: 6a0c1210fbc373592c01caed. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### ReMA Accuracy: 17.62. Result ID: 6a0c1210fbc373592c01caeb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Method Category":"Agentic Method","Backbone":"GPT-5"}. This record does not assert independent reproduction. ### VAM Accuracy: 17.11. Result ID: 6a0c1210fbc373592c01caee. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Category":"Agentic Method"}. This record does not assert independent reproduction. ### ReMA Accuracy: 14.23. Result ID: 6a0c1210fbc373592c01caec. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Method Category":"Agentic Method","Backbone":"Qwen3VL-A22B"}. This record does not assert independent reproduction. ### GPT-5 Accuracy: 10.15. Result ID: 6a0c1210fbc373592c01cae4. [Reporting paper]() Recorded evaluation configuration: {"Method Category":"End-to-End MLLM"}. This record does not assert independent reproduction. ### Qwen3-VL-235B-A22B Accuracy: 9.09. Result ID: 6a0c1210fbc373592c01cae5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Category":"End-to-End MLLM"}. This record does not assert independent reproduction. ### Qwen3-VL-30B-A3B Accuracy: 8.33. Result ID: 6a0c1210fbc373592c01cae6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Category":"End-to-End MLLM"}. This record does not assert independent reproduction. ### Nemotron-v2-12B Accuracy: 7.52. Result ID: 6a0c1210fbc373592c01cae9. [Reporting paper]() Recorded evaluation configuration: {"Method Category":"End-to-End MLLM"}. This record does not assert independent reproduction. ### Video-XL-2-8B Accuracy: 6.02. Result ID: 6a0c1210fbc373592c01cae7. [Reporting paper]() Recorded evaluation configuration: {"Method Category":"End-to-End MLLM"}. This record does not assert independent reproduction. ### LongVT-7B Accuracy: 5.83. Result ID: 6a0c1210fbc373592c01caea. [Reporting paper]() Recorded evaluation configuration: {"Method Category":"Agentic Method"}. This record does not assert independent reproduction. ### Eagle-2.5-8B Accuracy: 3.76. Result ID: 6a0c1210fbc373592c01cae8. [Reporting paper]() Recorded evaluation configuration: {"Method Category":"End-to-End MLLM"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.