MirrorBench: Evaluating Self-centric Intelligence in MLLMs by Introducing a Mirror

TL;DR

MirrorBench is a new benchmark inspired by psychological mirror tests, designed to evaluate self-centric intelligence in multimodal large language models through progressively challenging tasks.

Contribution

It introduces a systematic, psychology-inspired framework for assessing self-referential understanding in embodied MLLMs, revealing current limitations.

Findings

MLLMs perform poorly on self-referential tasks compared to humans.

The benchmark reveals fundamental gaps in self-awareness in current models.

Abstract

Recent progress in Multimodal Large Language Models (MLLMs) has demonstrated remarkable advances in perception and reasoning, suggesting their potential for embodied intelligence. While recent studies have evaluated embodied MLLMs in interactive settings, current benchmarks mainly target capabilities to perceive, understand, and interact with external objects, lacking a systematic evaluation of self-centric intelligence. To address this, we introduce MirrorBench, a simulation-based benchmark inspired by the classical Mirror Self-Recognition (MSR) test in psychology. MirrorBench extends this paradigm to embodied MLLMs through a tiered framework of progressively challenging tasks, assessing agents from basic visual perception to high-level self-representation. Experiments on leading MLLMs show that even at the lowest level, their performance remains substantially inferior to human performance, revealing fundamental limitations in self-referential understanding. Our study bridges psychological paradigms and embodied intelligence, offering a principled framework for evaluating the emergence of general intelligence in large models. Project page: https://fflahm.github.io/mirror-bench-page/.