Research
WorkFactory builds executable professional tasks to train WorkPilot agent models
Researchers built 112,686 tasks across 612 occupations and fine-tuned Qwen 4B, 27B and 35B-A3B models into the WorkPilot family. WorkPilot-35B-A3B rose from 879 to 1565 Elo on GDPval-AA v2.1.
Every outlet that ran it 1 article
Hugging Face PapersResearchLead
WorkFactory: Turning Professional Intents into Executable Work for Scalable Agent Training