ResearchBenchmarksVideo Question Answering on EgoSchema 2023 (test)Follow58.55AccuracyGSMT-CLIP33.912440.308746.70553.1013Jan 31, 2026Evaluation ResultsMethodMethodLinksAccuracyGSMT-CLIPPre-trained model=CLIP...Pre-trained model=CLIP (ViT/B-32)2026.0158.55GSMT-AIOPre-trained model=All-...Pre-trained model=All-in-One-Base2026.0158.28MIST-CLIPPre-trained model=CLIP...Pre-trained model=CLIP (ViT/B-32)2026.0156.42MIST-AIOPre-trained model=All-...Pre-trained model=All-in-One-Base2026.0156.27VideoReCap2026.0150.23EgoVLP2026.0134.86