VGenST-Bench
Explore score history ↗VGenST-Bench evaluates spatio-temporal reasoning in multimodal large language models using 1,200 procedurally generated videos with controlled scene composition, camera trajectory, and reasoning targets. It covers 12 reasoning tasks across three spatial scales and 12 QA types across three reasoning levels, with multiple choice and open-ended question variants.
- Released
- 2026-05-21
- Evaluates
- General AI, Reasoning, Vision & 3D
- Openness
- unknown
- Importer
- Claire Radar
- Review status
- unreviewed
VGenST-Bench paper and code
No reported scores are on record for this benchmark yet.