2026-10-22 · America/Los_Angeles · 论文 · #1

StreamEval: Evaluating Interactive Streaming Video Generation Models

As streaming video generation advances, enabling real-time interaction with generated video is becoming increasingly important. We introduce StreamEval, a unified benchmark for interactive streaming video generation spanning five tasks: avatar dialogue, avatar performance, world generation, instruction editing, and reference editing. StreamEval comprises 300 scripted test cases lasting 15-300s, each specifying initial assets, incrementally arriving multimodal requests, and explicit requirements for state changes and content preservation. Performance is assessed along four evaluation…

热度 65.3 / 100；排名与评分保留该期记录。