ScreenSpot-Pro
A GUI grounding benchmark that measures how accurately a multimodal model can locate a referenced UI element (return its position) given a natural-language instruction and a full-screen, high-resolution screenshot of professional desktop software across 23 applications, 5 industries, and 3 operating systems.
What this benchmark measures
A GUI grounding benchmark that measures how accurately a multimodal model can locate a referenced UI element (return its position) given a natural-language instruction and a full-screen, high-resolution screenshot of professional desktop software across 23 applications, 5 industries, and 3 operating systems.
Rows on this page are sourced from public benchmark artifacts, leaderboard exports, or source-linked model reports. Each row keeps benchmark version, source model name, and available run details attached to the score.
The metric shown here is accuracy. It should be interpreted within ScreenSpot-Pro, not compared as part of a site-wide ranking.
Frequently asked
What is ScreenSpot-Pro?
A GUI grounding benchmark that measures how accurately a multimodal model can locate a referenced UI element (return its position) given a natural-language instruction and a full-screen, high-resolution screenshot of professional desktop software across 23 applications, 5 industries, and 3 operating systems. It is a multimodal benchmark measured by accuracy.
What does accuracy mean on ScreenSpot-Pro?
ScreenSpot-Pro reports accuracy (%); higher is better. Scores are shown only within ScreenSpot-Pro and are never averaged with other benchmarks.
What is the top reported ScreenSpot-Pro score?
Claude Opus 4.8 has the top reported score on ScreenSpot-Pro: 87.9% (accuracy).
Why do ScreenSpot-Pro scores differ across runs?
Harness, scaffold, reasoning effort, and prompt setup change results, so two runs of the same model can differ. evals.report keeps each score with its run context so the differences stay visible.
Does evals.report rank models across benchmarks?
No. ScreenSpot-Pro scores are shown within their own metric; evals.report never combines benchmarks into a composite ranking or a single "best model".