VisualAgentBench: Towards Large Multimodal Models as Visual Foundation Agents
This provides a foundational benchmark for researchers and developers working on LMMs to advance visual foundation agents, though it is incremental as it builds on existing LMM frameworks.
The authors tackled the lack of challenging benchmarks for Large Multimodal Models (LMMs) as visual foundation agents by introducing VisualAgentBench (VAB), a comprehensive benchmark that tests LMMs across diverse scenarios like Embodied and Graphical User Interface tasks, demonstrating their developing capabilities through evaluations of 17 models and showing performance improvements via behavior cloning.
Large Multimodal Models (LMMs) have ushered in a new era in artificial intelligence, merging capabilities in both language and vision to form highly capable Visual Foundation Agents. These agents are postulated to excel across a myriad of tasks, potentially approaching general artificial intelligence. However, existing benchmarks fail to sufficiently challenge or showcase the full potential of LMMs in complex, real-world environments. To address this gap, we introduce VisualAgentBench (VAB), a comprehensive and pioneering benchmark specifically designed to train and evaluate LMMs as visual foundation agents across diverse scenarios, including Embodied, Graphical User Interface, and Visual Design, with tasks formulated to probe the depth of LMMs' understanding and interaction capabilities. Through rigorous testing across nine proprietary LMM APIs and eight open models, we demonstrate the considerable yet still developing agent capabilities of these models. Additionally, VAB constructs a trajectory training set constructed through hybrid methods including Program-based Solvers, LMM Agent Bootstrapping, and Human Demonstrations, promoting substantial performance improvements in LMMs through behavior cloning. Our work not only aims to benchmark existing models but also provides a solid foundation for future development into visual foundation agents. Code, train \& test data, and part of fine-tuned open LMMs are available at \url{https://github.com/THUDM/VisualAgentBench}.