InteractionBench: A Real-Time Interaction Benchmark for Streaming Video Systems

arXiv:2610.05775 · cs.CV · Submitted 2026-10-05 · Read on arXiv

cs.CV

Submitted: 2026-10-05

Updated: 2026-10-05

Comments: Project page: https://www.enxinsong.com/projects/interactionbench/ Code: https://github.com/Espere-1119-Song/InteractionBench Data: https://huggingface.co/datasets/InteractionBench/InteractionBench

Code: https://github.com/Espere-1119-Song/InteractionBench

Project page: https://www.enxinsong.com/projects/interactionbench

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Benchmark — 1× in this paper · explained in 2 episodes

Sources

• Qwen3-VL Technical Report
• Flash-VStream: Memory-Based Real-Time Understanding for Long Video Streams
• JoyAI-VL-Interaction: Real-Time Vision-Language Interaction Intelligence
• ProactiveVideoQA: A Comprehensive Benchmark Evaluating Proactive Interactions in Video Large Language Models
• RIVER: A Real-Time Interaction Benchmark for Video LLMs
• OmniPro: A Comprehensive Benchmark for Omni-Proactive Streaming Video Understanding
• OmniInteract: Benchmarking Real-World Streaming Interaction for Real-Time Omnimodal Assistants
• Don't Pause! Every prediction matters in a streaming video
• LiveProBench: Can Streaming Video Models Really Interact Like Humans?
• Eyes Wide Open: Ego Proactive Video-LLM for Streaming Video
• IPIBench: Evaluating Interactive Proactive Intelligence of MLLMs under Continuous Streams
• PaSBench-Video: A Streaming Video Benchmark for Proactive Safety Warning
• MOSS-Video-Preview: Toward Real-Time Video Understanding via Cross-Attention
• VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding
• Qwen2.5-Omni Technical Report
• LLaVA-OneVision: Easy Visual Task Transfer
• MiniCPM-o 4.5: Towards Real-Time Full-Duplex Omni-Modal Interaction
• LLaVA-OneVision-2: Towards Next-Generation Perceptual Intelligence
• Qwen3 Technical Report
• gpt-oss-120b & gpt-oss-20b Model Card

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models