How Stale Is Your AI?

Two dates decide how current a model really is: the day the lab shipped it, and the day it stopped reading. This page puts both on one shelf for 20 models across 8 labs and counts upward from each one live. Ten of the twenty have a training cutoff their lab actually publishes, and every model name links to the lab document the date came from. The same table is served as models.json, next to a block of text you can paste into an AGENTS.md so an agent checks that file instead of trusting its own memory of which model is current.