ShotBench: Expert-Level Cinematic Understanding in Vision-Language Models Paper • 2506.21356 • Published Jun 26, 2025 • 22
