EventSTU: Event-Guided Efficient Spatio-Temporal Understanding for Video-LLMs

Benchmark Model Rank Results
video-question-answering-on-mlvuEventSTU (Qwen3-VL-8B)–Accuracy: 76.49