SOP-Bench: A new benchmark for evaluating AI agents on real business procedures

Amazon Science Version 1 original current

Imported from official source

Extendable framework enables testing agents on the full set of capabilities required to successfully complete a procedure, not isolated proxy tasks.

This version

Version
1 of 1
Recorded
September 20, 2026 19:52
Change
Initial
Content hash
fc14e866bbb95a4b079bbf3a84e68381
All versions
Revision history

Officially records where a publication came from, not whether it is true. Imported records are reproduced from an organization's own official source.