# Referring Video Object Segmentation on MeViS (val-u) Source: [Referring Video Object Segmentation on MeViS (val-u)]() Collection index: [llms.txt]() Data updated: 2026-03-31T11:14:11.823Z Evaluation of referring video object segmentation performance on the 'val-u' subset of the MeViS dataset, reporting metrics such as J&F, F1, and HOTA. 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 | Referring Video Object Segmentation | | Dataset (including any stated split) | MeViS (val-u) | | Metric | J&F Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 17 of 17 | | 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 - [J&F Score](): Higher is better; unit: PERCENT. - [F1 Score](): Higher is better; unit: PERCENT. - [HOTA](): Higher is better; unit: PERCENT. ## Recorded results | Method | J&F Score | Result date | Report paper | | --- | --- | --- | --- | | MolmoPoint-8B | 72.2 | Not recorded | 69cb553d0636e5bf4afca121 | | Molmo2-8B | 70.8 | Not recorded | 69cb553d0636e5bf4afca121 | | Molmo2-4B | 70 | Not recorded | 69cb553d0636e5bf4afca121 | | Molmo2-O-7B | 69.7 | Not recorded | 69cb553d0636e5bf4afca121 | | Sa2VA-Qwen3-VL-4B | 57.1 | Not recorded | 69cb553d0636e5bf4afca121 | | Sa2VA-8B | 57 | Not recorded | 69cb553d0636e5bf4afca121 | | VideoMolmo-7B | 57 | Not recorded | 69cb553d0636e5bf4afca121 | | VideoLISA | 53.2 | Not recorded | 69cb553d0636e5bf4afca121 | | Gemini 2.5 Pro | 52.8 | Not recorded | 69cb553d0636e5bf4afca121 | | Molmo + SAM 2 | 51.5 | Not recorded | 69cb553d0636e5bf4afca121 | | Gemini 3 Pro | 51.1 | Not recorded | 69cb553d0636e5bf4afca121 | | VideoGLaMM | 50.6 | Not recorded | 69cb553d0636e5bf4afca121 | | Qwen3-VL-8B | 34.4 | Not recorded | 69cb553d0636e5bf4afca121 | | Gemini 2.5 Flash | 31.8 | Not recorded | 69cb553d0636e5bf4afca121 | | Qwen3-VL-4B | 30.6 | Not recorded | 69cb553d0636e5bf4afca121 | | GPT-5 | 26.5 | Not recorded | 69cb553d0636e5bf4afca121 | | GPT-5 mini | 15.4 | Not recorded | 69cb553d0636e5bf4afca121 | ### MolmoPoint-8B J&F Score: 72.2. Result ID: 69cbac6e47f34ce924b4c9bd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"MolmoPoint"}. This record does not assert independent reproduction. ### Molmo2-8B J&F Score: 70.8. Result ID: 69cbac6e47f34ce924b4c9bb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Fully open"}. This record does not assert independent reproduction. ### Molmo2-4B J&F Score: 70. Result ID: 69cbac6e47f34ce924b4c9ba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Fully open"}. This record does not assert independent reproduction. ### Molmo2-O-7B J&F Score: 69.7. Result ID: 69cbac6e47f34ce924b4c9bc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Fully open"}. This record does not assert independent reproduction. ### Sa2VA-Qwen3-VL-4B J&F Score: 57.1. Result ID: 69cbac6e47f34ce924b4c9b7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Specialized open models"}. This record does not assert independent reproduction. ### Sa2VA-8B J&F Score: 57. Result ID: 69cbac6e47f34ce924b4c9b6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Specialized open models"}. This record does not assert independent reproduction. ### VideoMolmo-7B J&F Score: 57. Result ID: 69cbac6e47f34ce924b4c9b9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Fully open"}. This record does not assert independent reproduction. ### VideoLISA J&F Score: 53.2. Result ID: 69cbac6e47f34ce924b4c9b4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Category":"Specialized open models"}. This record does not assert independent reproduction. ### Gemini 2.5 Pro J&F Score: 52.8. Result ID: 69cbac6e47f34ce924b4c9b0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"API call only"}. This record does not assert independent reproduction. ### Molmo + SAM 2 J&F Score: 51.5. Result ID: 69cbac6e47f34ce924b4c9b8. [Reporting paper]() Recorded evaluation configuration: {"Category":"Fully open"}. This record does not assert independent reproduction. ### Gemini 3 Pro J&F Score: 51.1. Result ID: 69cbac6e47f34ce924b4c9af. [Reporting paper]() Recorded evaluation configuration: {"Category":"API call only"}. This record does not assert independent reproduction. ### VideoGLaMM J&F Score: 50.6. Result ID: 69cbac6e47f34ce924b4c9b5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Category":"Specialized open models"}. This record does not assert independent reproduction. ### Qwen3-VL-8B J&F Score: 34.4. Result ID: 69cbac6e47f34ce924b4c9b3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Open weights only"}. This record does not assert independent reproduction. ### Gemini 2.5 Flash J&F Score: 31.8. Result ID: 69cbac6e47f34ce924b4c9b1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"API call only"}. This record does not assert independent reproduction. ### Qwen3-VL-4B J&F Score: 30.6. Result ID: 69cbac6e47f34ce924b4c9b2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Open weights only"}. This record does not assert independent reproduction. ### GPT-5 J&F Score: 26.5. Result ID: 69cbac6e47f34ce924b4c9ad. [Reporting paper]() Recorded evaluation configuration: {"Category":"API call only"}. This record does not assert independent reproduction. ### GPT-5 mini J&F Score: 15.4. Result ID: 69cbac6e47f34ce924b4c9ae. [Reporting paper]() Recorded evaluation configuration: {"Category":"API call only"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.