Benchmark RadarAI BENCHMARK PROFILE
IPIBench
Benchmark for interactive proactive intelligence of MLLMs under streaming video, covering proactive monitoring, task management, and interleaved requests.
- Released
- 2026-05-26
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
Existing benchmarks overlook dynamic multi-turn proactive interactions. IPIBench fills this gap for streaming assistant evaluation.
Motivation
Recent multimodal large language models (MLLMs) achieve strong performance on reactive question answering, but real-world streaming assistants require proactive reasoning over continuous visual inputs.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.