Social IQa

Sap et al., 2019

What it measures

A published test we don't mirror: it supplies a reference figure, not test items

Social IQa is 38,000 multiple-choice questions about everyday social situations: why someone did something, how they feel, what they will do next.

An example from the paper: Jordan wanted to tell Tracy a secret, so Jordan leaned towards Tracy. Why? To make sure no one else could hear.

Written for this page in the style of the test, not one of its questions, so the test set stays unspoiled

After the favourspecimen · not scored

Pick the most likely answer.

(a) Situation

Riley stayed late to help Sam finish the report, even though Riley had plans that evening.

(b) Question

How would Sam feel as a result?

Figure 1. A three-way question about how people react to an everyday situation. In the style of Social IQa. Answer: (a).

The faculty-level figures we take from it, 0–100

Humans

84.4

Social IQa (Sap et al., 2019)

Best published AI

80.0

CogToM, GPT-5.1 (a different test)

From another test: CogToM

These are third-party results on different questions from ours. Where the numbers come from