VisualToolBench is the first benchmark designed to evaluate how well Multimodal Large Language Models (MLLMs) can dynamically interact with and reason about visual information. It shifts the paradigm from passively “thinking about images” to actively “thinking with images,” treating images as a manipulable cognitive workspace.
Multimodal Large Language Models (MLLMs) are increasingly applied in realworld scenarios where user-provided images are often imperfect, requiring active image manipulations such as cropping, editing, or enhancement to uncover salient visual cues. Beyond static visual perception, MLLMs must also think with images: dynamically transforming visual content and integrating it with other tools to solve complex tasks. However, this shift from treating vision as passive context to a manipulable cognitive workspace remains underexplored. Most existing benchmarks still follow a think about images paradigm, where images are regarded as static inputs.
To address this gap, we introduce VisualToolBench (VTB), a benchmark that evaluates MLLMs’ ability to perceive, transform, and reason across complex visual–textual tasks under the think-with-images paradigm. VTB comprises 1,204 challenging, open-ended vision tasks (603 single-turn, 601 multi-turn) spanning across five diverse domains, each paired with detailed rubrics to enable systematic evaluation. Our evaluation shows that current MLLMs struggle with tasks requiring effective integration of vision and general-purpose tools. Even the strongest model (GPT-5-think) reaches only 18.68% pass rate. We further observe divergent tool-use behaviors, with OpenAI models benefiting from diverse image manipulations while Gemini-2.5-pro shows no improvement. By introducing the first benchmark centered on think with images, VTB offers critical insights for advancing visual intelligence in MLLMs.