ScratchEval: Are GPT-4o Smarter than My Child? Evaluating Large Multimodal Models with Visual Programming Challenges

TL;DR

ScratchEval is a new benchmark that assesses large multimodal models' ability to understand and reason about visual programming tasks using Scratch, addressing limitations of previous image-to-code evaluations.

Contribution

The paper introduces ScratchEval, a comprehensive benchmark for evaluating LMMs' visual programming reasoning using Scratch, combining visual understanding with code logic.

Findings

LMMs struggle with integrated visual and logical reasoning tasks.

ScratchEval reveals gaps in current multimodal models' programming understanding.

Benchmark encourages development of models with better logical and visual integration capabilities.

Abstract

Recent advancements in large multimodal models (LMMs) have showcased impressive code generation capabilities, primarily evaluated through image-to-code benchmarks. However, these benchmarks are limited to specific visual programming scenarios where the logic reasoning and the multimodal understanding capacities are split apart. To fill this gap, we propose ScratchEval, a novel benchmark designed to evaluate the visual programming reasoning ability of LMMs. ScratchEval is based on Scratch, a block-based visual programming language widely used in children's programming education. By integrating visual elements and embedded programming logic, ScratchEval requires the model to process both visual information and code structure, thereby comprehensively evaluating its programming intent understanding ability. Our evaluation approach goes beyond the traditional image-to-code mapping and focuses on unified logical thinking and problem-solving abilities, providing a more comprehensive and challenging framework for evaluating the visual programming ability of LMMs. ScratchEval not only fills the gap in existing evaluation methods, but also provides new insights for the future development of LMMs in the field of visual programming. Our benchmark can be accessed at https://github.com/HKBUNLP/ScratchEval .