claude-deprecation-vs-retirement-distinction

IN premise — entries/2026/06/21/wiki-Claude_language_model-chunk-2.md

Created 2026-06-21T09:50:09+00:00

Anthropic distinguishes between model deprecation (still accessible with restrictions) and retirement (fully removed from service).

Summary

Anthropic treats its models as having two distinct end-of-life stages: deprecated, meaning they still respond but carry restrictions or warnings, and retired, meaning they have been fully removed and will no longer accept requests. This distinction matters for the system because a deprecated model should still be treated as operational and routable, while a retired one must be handled as permanently unavailable, so the system knows when to fall back or migrate rather than assume a call has simply failed.

Dependents

These beliefs depend on this one:

IN anthropic-has-comprehensive-model-lifecycle-policy — Anthropic maintains a structured model lifecycle that distinguishes between deprecation (models still accessible with restrictions) and retirement (scheduled removal from service), with at least one model scheduled for retirement in August 2026. Deprecated models remain available to paid subscribers and via API by request. Anthropic also preserves weights of retired models as a precaution in case models have morally relevant preferences, conducting exit interviews before retirement.