A new 2,013-sample Linux desktop benchmark shows frontier VLMs are far from solving temporal ordering and single-action reconstruction, with best ordering exact-match near 65%.
arXiv:2606.15673
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Desktop-Delta Bench: Do Computer-Use Models Understand Desktop GUI Transitions?
A new 2,013-sample Linux desktop benchmark shows frontier VLMs are far from solving temporal ordering and single-action reconstruction, with best ordering exact-match near 65%.