LLM Stats (Social IQa): leaderboard
Social IQa — social commonsense reasoning MCQ, as independently run by LLM-Stats.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around January 2027. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Phi-3.5-MoE-instruct | 78 | #700 |
| 2 | Phi-3.5-mini-instruct | 74.7 | #1167 |
| 3 | Phi-4 Mini Instruct | 72.5 | #1245 |
| 4 | Gemma 2 27B | 53.7 | #806 |
| 5 | Gemma 2 9B | 53.4 | #958 |
| 6 | Gemma 3n E4B | 50 | #878 |
| 7 | Gemma 3n E2B | 48.8 | #1158 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-social-iqa · How It Works · Data refreshed daily, snapshot 2026-10-20.