Beyond Seeing: Evaluating Multimodal LLMs on Tool-Enabled Image Perception, Transformation, and Reasoning
This addresses a gap in benchmarking for MLLMs in real-world scenarios where images need transformation and reasoning, offering insights for advancing visual intelligence, though it is incremental as it focuses on evaluation rather than new model development.
The paper tackles the problem of evaluating multimodal large language models (MLLMs) on tasks that require active image manipulation and tool integration, rather than just passive perception, by introducing VisualToolBench, a benchmark with 1,204 tasks across five domains, and finds that current models struggle, with the strongest achieving only an 18.68% pass rate.
Multimodal Large Language Models (MLLMs) are increasingly applied in real-world 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, a visual tool-use reasoning benchmark that rigorously evaluates MLLMs' ability to perceive, transform, and reason across complex visual-textual tasks under the think-with-images paradigm. VisualToolBench 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, VisualToolBench offers critical insights for advancing visual intelligence in MLLMs.