MEMO: Multi-Level Entity-Aware Memory for Streaming Video Understanding

Benchmark Model Rank Results
video-question-answering-on-streamingbenchQwen3-VL-8B + MEMO#5Real-Time Visual Understanding: 83.7
video-question-answering-on-streamingbenchQwen2.5-VL-7B + MEMO#10Real-Time Visual Understanding: 78.5
video-question-answering-on-streamingbenchLLaVA-OneVision-7B + MEMO#11Real-Time Visual Understanding: 73.2
video-question-answering-on-streamingbenchLLaVA-OneVision-0.5B + MEMO#12Real-Time Visual Understanding: 59.5