Event-VStream: Event-Driven Real-Time Understanding for Long Video Streams

TL;DR

Event-VStream introduces an event-driven framework that detects meaningful state transitions in long videos, enabling efficient, long-horizon reasoning and real-time understanding with low latency, outperforming existing methods.

Contribution

The paper presents a novel event-aware system that models videos as sequences of discrete events, improving real-time understanding and reasoning in long video streams compared to prior fixed-interval approaches.

Findings

+10.4 points on OVOBench-Realtime

Close performance to Flash-VStream-7B with LLaMA-3-8B backbone

Around 70% GPT-5 win rate on 2-hour Ego4D streams

Abstract

Real-time understanding of long video streams remains challenging for multimodal large language models (VLMs) due to redundant frame processing and rapid forgetting of past context. Existing streaming systems rely on fixed-interval decoding or cache pruning, which either produce repetitive outputs or discard crucial temporal information. We introduce Event-VStream, an event-aware framework that represents continuous video as a sequence of discrete, semantically coherent events. Our system detects meaningful state transitions by integrating motion, semantic, and predictive cues, and triggers language generation only at those boundaries. Each event embedding is consolidated into a persistent memory bank, enabling long-horizon reasoning while maintaining low latency. Across OVOBench-Realtime, and long-form Ego4D evaluations, Event-VStream achieves competitive performance. It improves over a VideoLLM-Online-8B baseline by +10.4 points on OVOBench-Realtime, achieves performance close to Flash-VStream-7B despite using only a general-purpose LLaMA-3-8B text backbone, and maintains around 70% GPT-5 win rate on 2-hour Ego4D streams.