Principia Benchmark Exposes Video Models' Poor Physics Understanding

Researchers from IISc and Johns Hopkins introduced Principia, a benchmark of relational physics tests across about 500 real-world scenarios, finding leading video models score high on VBench (0.8) but fail physics consistency (0.42).

2026-09-07 ~ 2026-09-07 · 3 related posts

• Principia benchmark: video models score 0.8 on VBench but under 0.42 on physics consistency — CSProfKGD · 2026-09-07
• Principia benchmark: 500+ real scenes show video models know less high-school physics than you'd think — anand_bhattad · 2026-09-07

1 near-duplicate retellings: anand_bhattad