IPIBench - Task Cancellation: 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 task cancellation (stop responding after the user cancels the monitoring task); percent, higher is better.

Metric: Accuracy (%) on task cancellation (stop responding after the user cancels the monitoring task), 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 May 2027. 15 models tracked. Updated 14 Nov 2026.

Top models

#ModelScoreOverall rank
1GPT-5.438.43#135
2Gemini 2.5 Pro24.54#276
3Gemini 3 Pro19.44#145
4GLM-4.6V15.28#710
5MOSS-VL-Realtime15.28
6JoyAI-VL-Interaction14.81
7Qwen 3.5 Plus12.96#242
8GPT-4o11.57#751
9VideoLLM-online-8B7.87
10InternVL3-8B6.94#1782
11Flash-VStream-7B2.31
12LLaVA-OneVision-7B1.39#2981
13Qwen 3 VL 8B1.39#917
14Qwen 3.5 9B (Harness: AURA2 IPIBench offline30frames; Thinking: Off)0.46
15Qwen 3.5 9B (Harness: AURA2 IPIBench offline30frames; Thinking: On)0

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=ipibench-task-cancellation · How It Works · Data refreshed daily, snapshot 2026-11-21.