Physical Plausibility Verification on PhysicsIQ CogVideoX-5B I2V (N=16) benchmark leaderboard · SOTA2 Research

Physical Plausibility Verification on PhysicsIQ CogVideoX-5B I2V (N=16)

0.302Spatial IoUOracle (Upper bound)

Evaluation Results

Method	Links
Oracle (Upper bound)2026.06		0.302	0.246	0.162	0.007	41.3
WMReward2026.06		0.235	0.21	0.13	0.007	33.2
Qwen3-VL2026.06		0.227	0.193	0.117	0.007	30.9
GEOPHYS2026.06		0.225	0.232	0.128	0.006	34.1
Qwen2.5-VL2026.06		0.168	0.132	0.072	0.008	21.3
Baseline2026.06		0.143	0.133	0.07	0.008	19.9
VideoMAE2026.06		0.122	0.044	0.042	0.008	11.4