StreamOPD: A Post-Training Recipe with Spatio-Temporal Cue Gating for Streaming Video Understanding
Open paper
Benchmark
Model
Rank
Results
video-question-answering-on-streamingbench
ST-CueGate (Qwen3.5-4B, n=4) (extra data)
#4
Real-Time Visual Understanding: 84.55
Rank counts only results with a code link.