Video Question Answering on EgoPlan-Bench
43.06AccuracyGPT-5.2
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5.2Encoding=Uniform-8, Variant=Plain2026.06 | 43.06 | |
| Gemini 3 FlashEncoding=Uniform-8, Variant=Plain2026.06 | 37.96 | |
| Gemini 3 FlashEncoding=Cascade, Variant=Seed2026.06 | 30.8 | |
| FullEvo (Cat.)Backbone=Gemini 3 Flash, Encoding=Cascade2026.06 | 30.04 | |
| Gemini 3 FlashEncoding=Cascade, Variant=+Evolve2026.06 | 29.93 | |
| FullEvo (Guide)Backbone=GPT-5.2, Encoding=Cascade2026.06 | 29.39 | |
| FullEvo (Guide)Backbone=Gemini 3 Flash, Encoding=Cascade2026.06 | 28.85 | |
| FullEvo (Cat.)Backbone=GPT-5.2, Encoding=Cascade2026.06 | 28.85 | |
| GPT-5.2Encoding=Cascade, Variant=Seed2026.06 | 28.74 | |
| GPT-5.2Encoding=Cascade, Variant=Plain2026.06 | 28.42 | |
| Gemini 3 FlashEncoding=Cascade, Variant=+SkillMemCat2026.06 | 28.31 | |
| GPT-5.2Encoding=Cascade, Variant=+Evolve2026.06 | 28.31 | |
| GPT-5.2Encoding=Cascade, Variant=+SkillMemCat2026.06 | 28.09 | |
| Gemini 3 FlashEncoding=Cascade, Variant=Plain2026.06 | 24.62 |