A capability-driven benchmark of 400 bilingual real-world tasks shows current proactive agents fail >50% of the time, with framework architecture impacting performance more than base model choice.
An ICS import preview, a screen where no calendar has been selected, or a dialog where import/save has not been clicked does not count as a created event
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
UniClawBench: A Universal Benchmark for Proactive Agents on Real-World Tasks
A capability-driven benchmark of 400 bilingual real-world tasks shows current proactive agents fail >50% of the time, with framework architecture impacting performance more than base model choice.