ScreenSpot-Pro: Results as Published

ScreenSpot-Pro tests whether a multimodal AI model can find the right spot on a screen. Each item gives a high-resolution screenshot of professional software, across 23 applications in five industries and three operating systems, plus a written instruction, and the model must point to the matching element. A prediction counts as correct if its predicted point falls inside the target element's bounding box.

Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.