Observed arrival · 2026-10-25

AutoFactory-Bench: Can AI run a factory?

autofactorybench.com Visit website

Editorial interest 78/100 Selection signal · not a rating of the site

A benchmark page compares AI agents writing PLC programs against tasks in a simulated factory, scoring both task completion and safety.

Landing page captured for the 2026-10-25 issue.

For: Researchers testing AI agents on industrial control tasks
Worth noticing: The page separates safe success from raw task completion and shows model comparisons with 95% intervals.

Field notes

Task examples range from coordinating two workpiece types across machining cells with a shared robot to maintaining tool retraction as work-stage pockets rotate. The page separates hard, medium, and easy cases, shows model and harness comparisons with 95% intervals, and distinguishes safe success from raw task completion; its displayed index also attributes Track 1 to PLCWorld (arXiv 2610.02982).

Observed signals

Read the marks

Editorial observations of this landing page, not a rating.

○OpenPublic substance visible

✦PrettyNotable craft visible

●ProPolished or operationally mature

◎NicheUnusually specific use

One card from the complete issue

Certified Dino Tamer

512,544 arrived 1,000 judged 976 catalogued Enter the complete issue