BabyVision: Visual Reasoning Beyond Language

TL;DR

BabyVision is a new benchmark designed to evaluate core visual reasoning abilities in multimodal language models independently of language, revealing significant gaps compared to human performance and highlighting the need for improved visual perception in AI.

Contribution

The paper introduces BabyVision, a comprehensive benchmark for assessing fundamental visual skills in multimodal models, and provides initial evaluations showing current models lag behind humans in basic visual reasoning.

Findings

Leading MLLMs perform significantly below human baseline.

Current models lack fundamental visual primitives despite language capabilities.

Progress in BabyVision advances toward human-level visual perception.

Abstract

While humans develop core visual skills long before acquiring language, contemporary Multimodal LLMs (MLLMs) still rely heavily on linguistic priors to compensate for their fragile visual understanding. We uncovered a crucial fact: state-of-the-art MLLMs consistently fail on basic visual tasks that humans, even 3-year-olds, can solve effortlessly. To systematically investigate this gap, we introduce BabyVision, a benchmark designed to assess core visual abilities independent of linguistic knowledge for MLLMs. BabyVision spans a wide range of tasks, with 388 items divided into 22 subclasses across four key categories. Empirical results and human evaluation reveal that leading MLLMs perform significantly below human baselines. Gemini3-Pro-Preview scores 49.7, lagging behind 6-year-old humans and falling well behind the average adult score of 94.1. These results show despite excelling in knowledge-heavy evaluations, current MLLMs still lack fundamental visual primitives. Progress in BabyVision represents a step toward human-level visual perception and reasoning capabilities. We also explore solving visual reasoning with generation models by proposing BabyVision-Gen and automatic evaluation toolkit. Our code and benchmark data are released at https://github.com/UniPat-AI/BabyVision for reproduction.