ConsumerSimBench - Emotion Keywords: leaderboard

ConsumerSimBench Emotion Keywords measures whether the comments a model writes about 1,553 real RedNote trending topics reproduce the emotion keywords that real consumers expressed (ConsumerSimBench, 2026). The score is the percent of the criteria in that family that the comments cover, judged by GPT-5.2; higher is better.

Metric: Criterion coverage (%) for T2 emotion keywords (the affective vocabulary of the real comments, matched semantically): the share of a topic's criteria that the model's generated comment set covers, each criterion judged pointwise by GPT-5.2 against its definition and examples, averaged over the 1,553 RedNote trending topics of ConsumerSimBench (23,122 criteria in all); the model writes 10 to 30 realistic comments per topic from a neutral event description, one call at temperature 0.7; higher is better. Source: arxiv.org. Saturation forecast: Around November 2027. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3.1 Pro (Preview)56.9#99
2Gemini 3 Flash53.1#175
3Grok 4.2052.5#301
4Doubao Seed 1.8Thinking51#352
5Qwen 3.5 397B A17B50.4#264
6ConsumerSimBench DeepSeek-V3 (checkpoint unspecified)49.2
7Kimi K2.549#266
8MiniMax-M2.540.6#600
9MiMo-V2.5-Pro39.7#284
10Claude Opus 4.637.7#113
11GPT-4o35.8#751
12GPT-5.235.1#185
13Claude Sonnet 4.632.6#173

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

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