EmoTrans - Emotion Change Detection: leaderboard

EmoTrans (2026) emotion change detection: for each target person in a real-world video clip, the model decides whether their emotion changes; accuracy (%) over 1,000 questions, higher is better.

Metric: Accuracy (%) on Emotion Change Detection (1,000 questions: whether each target person's emotion changes), on EmoTrans, built from 1,000 manually annotated real-world video clips (OpenHumanVid and CH-SIMS v2) over 12 scenarios with one or more target persons; zero-shot, video with its text (and audio where the model accepts it); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-5.475.4#135
2Gemini 3 Pro74.7#145
3Qwen 3.5 122B A10B74#346
4Qwen 2.5 VL 72B Instruct73.9#841
5Gemini 3 FlashNon-reasoning73.7#175
6Qwen 3.5 27B72.8#375
7Qwen 3 VL 32B Instruct71.5#606
8Qwen 3 VL 8B71.2#917
9GLM-4.6V70#710
10Gemini 2.5 FlashNon-reasoning69.8#463
11InternVL3.5-8B69.5#1604
12GLM-4.5V66.2#754
13InternVL3.5-14B65.3#1504
14Seed 2.0 Lite64.9#300
15Qwen 3 Omni 30B A3B62.9#847

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

Interactive version: theaggregate.ai/benchmark?slug=emotrans-emotion-change-detection · How It Works · Data refreshed daily, snapshot 2026-11-21.