The State of Production Agents
Loading the latest editionThe State of Production Agents
Loading the latest editionFree report
A quarterly read on what AI agents in production actually measure, drawn entirely from submissions to a free benchmark on this site.
No edition has been generated yet. The first one publishes once there are enough submissions for the figures to mean anything — a cohort needs 20 submissions on a metric before that metric appears at all, and no cell below that is ever shown.
Every figure will come from a query against real submissions. Nothing here will be modelled or estimated, which is why it has to wait for the data rather than launching with placeholder numbers.
Run the 15-metric benchmark and be counted in the first edition
One email when there is enough data to publish. No reminders in the meantime.