# Video Question Answering on EgoPlan-Bench Source: [Video Question Answering on EgoPlan-Bench]() Collection index: [llms.txt]() Data updated: 2026-06-16T09:47:23.151Z Evaluation of model performance on the EgoPlan-Bench dataset for Video Question Answering, reporting 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) | EgoPlan-Bench | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 14 of 14 | | 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 | | --- | --- | --- | --- | | GPT-5.2 | 43.06 | Not recorded | 6a30d906d33dfb23e783d1b7 | | Gemini 3 Flash | 37.96 | Not recorded | 6a30d906d33dfb23e783d1b7 | | Gemini 3 Flash | 30.8 | Not recorded | 6a30d906d33dfb23e783d1b7 | | FullEvo (Cat.) | 30.04 | Not recorded | 6a30d906d33dfb23e783d1b7 | | Gemini 3 Flash | 29.93 | Not recorded | 6a30d906d33dfb23e783d1b7 | | FullEvo (Guide) | 29.39 | Not recorded | 6a30d906d33dfb23e783d1b7 | | FullEvo (Guide) | 28.85 | Not recorded | 6a30d906d33dfb23e783d1b7 | | FullEvo (Cat.) | 28.85 | Not recorded | 6a30d906d33dfb23e783d1b7 | | GPT-5.2 | 28.74 | Not recorded | 6a30d906d33dfb23e783d1b7 | | GPT-5.2 | 28.42 | Not recorded | 6a30d906d33dfb23e783d1b7 | | Gemini 3 Flash | 28.31 | Not recorded | 6a30d906d33dfb23e783d1b7 | | GPT-5.2 | 28.31 | Not recorded | 6a30d906d33dfb23e783d1b7 | | GPT-5.2 | 28.09 | Not recorded | 6a30d906d33dfb23e783d1b7 | | Gemini 3 Flash | 24.62 | Not recorded | 6a30d906d33dfb23e783d1b7 | ### GPT-5.2 Accuracy: 43.06. Result ID: 6a311b96fbc373592c2de4ef. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Uniform-8","Variant":"Plain"}. This record does not assert independent reproduction. ### Gemini 3 Flash Accuracy: 37.96. Result ID: 6a311b96fbc373592c2de4e8. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Uniform-8","Variant":"Plain"}. This record does not assert independent reproduction. ### Gemini 3 Flash Accuracy: 30.8. Result ID: 6a311b96fbc373592c2de4e3. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Cascade","Variant":"Seed"}. This record does not assert independent reproduction. ### FullEvo (Cat.) Accuracy: 30.04. Result ID: 6a311b96fbc373592c2de4e6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Gemini 3 Flash","Encoding":"Cascade"}. This record does not assert independent reproduction. ### Gemini 3 Flash Accuracy: 29.93. Result ID: 6a311b96fbc373592c2de4e4. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Cascade","Variant":"+Evolve"}. This record does not assert independent reproduction. ### FullEvo (Guide) Accuracy: 29.39. Result ID: 6a311b96fbc373592c2de4ee. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"GPT-5.2","Encoding":"Cascade"}. This record does not assert independent reproduction. ### FullEvo (Guide) Accuracy: 28.85. Result ID: 6a311b96fbc373592c2de4e7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Gemini 3 Flash","Encoding":"Cascade"}. This record does not assert independent reproduction. ### FullEvo (Cat.) Accuracy: 28.85. Result ID: 6a311b96fbc373592c2de4ed. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"GPT-5.2","Encoding":"Cascade"}. This record does not assert independent reproduction. ### GPT-5.2 Accuracy: 28.74. Result ID: 6a311b96fbc373592c2de4ea. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Cascade","Variant":"Seed"}. This record does not assert independent reproduction. ### GPT-5.2 Accuracy: 28.42. Result ID: 6a311b96fbc373592c2de4e9. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Cascade","Variant":"Plain"}. This record does not assert independent reproduction. ### Gemini 3 Flash Accuracy: 28.31. Result ID: 6a311b96fbc373592c2de4e5. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Cascade","Variant":"+SkillMemCat"}. This record does not assert independent reproduction. ### GPT-5.2 Accuracy: 28.31. Result ID: 6a311b96fbc373592c2de4eb. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Cascade","Variant":"+Evolve"}. This record does not assert independent reproduction. ### GPT-5.2 Accuracy: 28.09. Result ID: 6a311b96fbc373592c2de4ec. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Cascade","Variant":"+SkillMemCat"}. This record does not assert independent reproduction. ### Gemini 3 Flash Accuracy: 24.62. Result ID: 6a311b96fbc373592c2de4e2. [Reporting paper]() Recorded evaluation configuration: {"Encoding":"Cascade","Variant":"Plain"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.