IPIBench - Reactive-after-Proactive: leaderboard

IPIBench (2026) streams videos to multimodal LLMs at one frame per second and checks whether they respond proactively at the right moment, manage monitoring tasks the user cancels or changes, and handle reactive questions interleaved with proactive tasks. This board is accuracy on reactive-after-proactive interaction (answer a query right after a proactive trigger); percent, higher is better.

Metric: Accuracy (%) on reactive-after-proactive interaction (answer a query right after a proactive trigger), on IPIBench (1,831 videos, 3,738 interaction instances from egocentric and exocentric datasets) under a unified streaming protocol at 1 FPS (offline models see a sliding window of the latest 16 frames); a proactive response counts only if the first trigger falls within one second of the ground-truth moment, multi-turn tasks only if every required trigger and answer is correct, reactive answers by bidirectional substring matching against up to 20 reference answers; higher is better. Source: arxiv.org. Saturation forecast: Around August 2028. 15 models tracked. Updated 14 Nov 2026.

Top models

#ModelScoreOverall rank
1GLM-4.6V39.37#721
2Qwen 3.5 9B (Harness: AURA2 IPIBench offline30frames; Thinking: On)33.16
3Qwen 3.5 Plus30.57#250
4GPT-5.429.28#138
5Gemini 2.5 Pro28.24#280
6Gemini 3 Pro26.69#146
7GPT-4o26.17#754
8Qwen 3 VL 8B23.05#927
9MOSS-VL-Realtime10.88
10InternVL3-8B9.33#1795
11LLaVA-OneVision-7B3.89#3002
12Qwen 3.5 9B (Harness: AURA2 IPIBench offline30frames; Thinking: Off)3.11
13JoyAI-VL-Interaction1.81
14VideoLLM-online-8B1.55
15Flash-VStream-7B0.78

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=ipibench-reactive-after-proactive · How It Works · Data refreshed daily, snapshot 2026-11-22.