A long-horizon hybrid-interface benchmark for CUA with 114 tasks across 8 real-world work domains, grounded in real user requests and publicly verifiable artifacts. Across frontier model-runtime pairings, the best PassRate reaches only 41.2%, showing the benchmark remains far from saturated.
Paper https://arxiv.org/abs/2606.09426
A long-horizon hybrid-interface benchmark for CUA with 114 tasks across 8 real-world work domains, grounded in real user requests and publicly verifiable artifacts. Across frontier model-runtime pairings, the best PassRate reaches only 41.2%, showing the benchmark remains far from saturated.
No Digg Deeper questions have been answered for this story yet.