OSWorld-Science: A Benchmark of Computer Use Agents for Learning and Using Scientific Software
OSWorld-Science benchmarks 12 VLMs on 146 scientific software tasks with artifact-based scoring.
OSWorld-Science is a benchmark and evaluation environment for computer-using agents based on visual language models performing scientific software workflows. It includes 12 VLMs and 146 expert-designed tasks spanning molecular drawing and retrosynthesis, pathology image analysis, statistical computing, and physical simulation. Execution-based evaluators inspect application state and artifacts such as molecular structures, segmentation masks, plots, and numerical results, awarding partial credit. Results show that even state-of-the-art VLMs with a strong harness still struggle on key scientific questions.
- 146 tasks across molecular, pathology, statistics, and simulation software.
- Evaluators score artifacts and application state with partial credit.
- Current VLMs still fail key scientific workflow questions.
- Analysis covers multilingual settings, reasoning effort, and context length.
Full article228 words · extracted from huggingface.co · click to collapse
Scientific software presents a demanding test for computer-using agents based on visual language models (VLMs): completing a research workflow requires interpreting specialized interfaces, manipulating scientific objects, and producing verifiable results. We thus introduce OSWorld-Science, a benchmark and evaluation environment that combines scientifically meaningful tasks, artifact-based evaluation, and an efficient agent harness for studying computer use in the scientific domain. The benchmark contains 12 VLMs and 146 high-quality tasks across several scientific domains and software configurations, covering workflows such as molecular drawing and retrosynthesis, pathology image analysis, statistical computing, and physical simulation. Tasks are developed through expert proposals and iterative human--AI co-design, with selection guided by scientific value and difficulty. Task-specific execution-based evaluators inspect application states and generated artifacts, including molecular structures, segmentation masks, plots, and numerical results, and award partial credit for incomplete outcomes. Our special harness integrates model adapters, interaction-loop control, and trajectory logging to support comparisons of models and interaction strategies. Our results show that current state-of-the-art VLMs with a strong harness still face challenges in addressing key questions in the scientific domains. We also analyze the benchmarking results across multi-linguistics, reasoning efforts, context length and other factors and derive several important conclusions and directions to assist future development. Overall, we provide an integrated framework connecting expert-defined scientific goals to verifiable software outcomes, enabling systematic evaluation of both agent capabilities and harness design in scientific workflows.
Text extracted automatically; images, tables and formatting may be missing. Original: https://huggingface.co/papers/2609.39903