psychology benchmark
Social IQa
The first large-scale benchmark for commonsense reasoning about social situations. Contains 38,000 multiple choice questions probing emotional and social intelligence in everyday situations, testing commonsense understanding of social interactions and theory of mind reasoning about the implied emotions and behavior of others.
Updated Aug 7, 2026
Published models9
Registry coverage9
MetricScore
EvidenceB
Social IQa leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | MI | 78.0% | 100.0% | 9 | C | |
| 2 | MI | 74.7% | 87.5% | 9 | C | |
| 3 | MI | 72.5% | 75.0% | 9 | C | |
| 4 | GO | 53.7% | 62.5% | 9 | C | |
| 5 | GO | 53.4% | 50.0% | 9 | C | |
| 6 | GO | 50.0% | 37.5% | 9 | C | |
| 7 | GO | 50.0% | 25.0% | 9 | C | |
| 8 | GO | 48.8% | 12.5% | 9 | C | |
| 9 | GO | 48.8% | 0.0% | 9 | C |
Score distribution
Top published rows on the benchmark's original scale.
Social IQa highlights
The top published results on this benchmark's own scale.
What is Social IQa?
Definition and scoring fields from the benchmark registry.
The first large-scale benchmark for commonsense reasoning about social situations. Contains 38,000 multiple choice questions probing emotional and social intelligence in everyday situations, testing commonsense understanding of social interactions and theory of mind reasoning about the implied emotions and behavior of others.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- Social IQa
- Modality
- text
- Primary category
- psychology
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- social-iqa|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about Social IQa.
Which model scores highest on Social IQa?
Phi-3.5-MoE-instruct is currently ranked first with 78.0%.
What does Social IQa measure?
The first large-scale benchmark for commonsense reasoning about social situations. Contains 38,000 multiple choice questions probing emotional and social intelligence in everyday situations, testing commonsense understanding of social interactions and theory of mind reasoning about the implied emotions and behavior of others.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
9 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.