New AtariBench Benchmark Tests Vision-Language Models' Spatiotemporal Reasoning in Gameplay
Researchers introduced AtariBench, a benchmark using Atari gameplay to evaluate how well vision-language models handle dynamic environments and spatiotemporal reasoning beyond static image tasks.
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-11-09 05:00 · arXiv cs.CV
AtariBench: Evaluating Spatiotemporal Reasoning of Vision-Language Models in Atari Gameplay