GlitchBench: Can Large Multimodal Models Detect Video Game Glitches?

Abstract

Large multimodal models (LMMs) have evolved from large language models (LLMs) to integrate multiple input modalities, such as visual inputs. This integration augments the capacity of LLMs for tasks requiring visual comprehen- sion and reasoning. However, the extent and limitations of their enhanced abilities are not fully understood, especially when it comes to real-world tasks. To address this gap, we introduce GlitchBench, a novel benchmark derived from video-game quality assurance tasks, to test and eval- uate the reasoning capabilities of LMMs. Our benchmark is curated from a variety of unusual and glitched scenarios from video games and aims to challenge both the visual and linguistic reasoning powers of LMMs in detecting and inter- preting out-of-the-ordinary events. Our evaluation shows that GlitchBench presents a new, interesting challenge to state-of-the-art LMMs. Code and data are available at: https://glitchbench.github.io/