ConsumerSimBench - Sentiment Flashpoints: leaderboard

ConsumerSimBench Sentiment Flashpoints measures whether the comments a model writes about 1,553 real RedNote trending topics reproduce the sentiment flashpoints 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 T1 sentiment flashpoints (the emotional triggers that pushed the topic onto the trending list): 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 October 2028. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 Flash39.5#175
2Gemini 3.1 Pro (Preview)38.9#99
3Grok 4.2036.8#301
4Qwen 3.5 397B A17B36.1#264
5Kimi K2.535.9#266
6Doubao Seed 1.8Thinking34.8#352
7ConsumerSimBench DeepSeek-V3 (checkpoint unspecified)33.9
8GPT-5.231.4#185
9MiniMax-M2.528.8#600
10MiMo-V2.5-Pro27.4#284
11Claude Opus 4.625.4#113
12Claude Sonnet 4.624.8#173
13GPT-4o23.5#751

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

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