GTM Bench scores LLMs and AI agents on real go-to-market work across two axes, how much of the job the system finishes and whether the output is groun ...