Fetching the paper…

OdysseyBench: Evaluating LLM Agents on Long-Horizon Complex Office Application Workflows · Around