mmWave-QA - Action Recognition (CoT): leaderboard

Metric: Accuracy (%; multiple-choice recognition of the performed action from mmWave radar point clouds serialized as text, 860 questions over the Phoenix, IWR6843 and IWR1443 radars and four body-region action categories; cot prompting, with an instruction to write a rationale before the answer). Source: arxiv.org. Saturation forecast: Around 2029. 4 models tracked.

Top models

#ModelScore
1Gemini 2.5 Pro32.79
2Gemini 2.5 Flash31.86
3GPT-530.23
4GPT-4o29.3

Interactive version: theaggregate.ai/benchmark?slug=mmwave-qa-action-recognition-cot · How It Works · Data refreshed daily, snapshot 2026-09-26.