What changed
OpenAI completed its 23 October shutdown of legacy GPT-3.5, GPT-4, GPT-4.1 nano, GPT-4o, o-series and first-generation image endpoints, including five fine-tuned model families. Requests now need the documented GPT-5.6 or GPT Image 2.5 replacements.
Why it matters
This is a production cutoff rather than a future notice: un-migrated workloads can fail outright, and replacement behavior, cost, latency and fine-tune portability can differ even when an alias looks similar.
What to do
Audit logs and configuration for every retired identifier, replay representative and denied cases on the documented replacement, compare output and cost, update fine-tune plans separately, and keep a tested rollback path that does not call a retired endpoint.
Caveat
The deprecation page establishes OpenAI-operated API availability, not every partner platform's schedule. Replacement models are not behaviorally equivalent, and a successful request does not establish task-level parity.