Open-o3-Video-v2: Seek, Ground, and Reason with Verifiable Spatio-Temporal Evidence

Benchmark Model Rank Results
video-question-answering-on-videommmuOpen-o3-Video-v2-8B#7Accuracy: 63.0
temporal-sentence-grounding-on-charades-timelensOpen-o3-Video-v2-8B#9mIoU: 51.1
video-question-answering-on-videoholmesOpen-o3-Video-v2-4B#10Accuracy: 48.0
audio-visual-question-answering-on-worldsenseOpen-o3-Video-v2-4B#14Accuracy: 46.3
video-question-answering-on-lvbenchOpen-o3-Video-v2-8B#27Accuracy: 53.9