OpenAI Kills 12 Model APIs Tomorrow (Oct 23) — Check Your Code Tonight

No affiliate links on this page — every link points to official docs or reporting, nothing sponsored.

Tomorrow morning, requests to o1, o4-mini, gpt-4-0613, and nine other OpenAI model IDs start failing. Not warning — failing. OpenAI announced this batch in April, the Sep 28 batch already died quietly, and this is the big one: the last of the GPT-3.5/GPT-4 names and the whole o-series mini lineup. Here's the list, the replacement map, and what to do tonight.

What's dying tomorrow

From OpenAI's deprecations page (the April 22 "Legacy GPT model snapshots" batch):

Three gotchas that catch people

• The bare aliases die too. Code calling o1 instead of o1-2024-12-17 is affected — OpenAI listed both, and the alias dies with the snapshot. This is the one that gets the cron jobs nobody remembers writing.
• Fine-tunes die with their base model. If you fine-tuned gpt-3.5-turbo in 2024, inference on that fine-tune stops tomorrow. Rebuilding takes real time — this is the item that doesn't fit in an evening, so check it first.
• gpt-4o-2024-05-13 is a GPT-4o snapshot, not GPT-4. If you pinned it for reproducibility — the responsible thing to do at the time — you're on this list and might not realise it.

What's NOT affected: ChatGPT itself, Azure OpenAI (different retirement calendar — check your Azure dates), and the newer 4o snapshots, which live until Jan 14, 2027.

The tonight checklist

• Grep every repo for the dead IDs. Not just o1 — grep the dated snapshots too. Include config files, .env.example files, Dockerfiles, and Terraform/Helm values.
• Check env vars and secrets managers. The model name often lives in an env var, not in code. Check your secrets manager, CI variables, and deployment platform's env panel.
• Check scheduled jobs and queues. A nightly batch job calling gpt-4-0613 fails silently and looks like "the queue is just slow" until someone checks the backlog.
• Check prompt-management configs. If you version prompts anywhere, the model ID is probably pinned there, not in your application code.
• Swap and smoke-test. Replace with the mapped model, then run one real request per changed call site. Parameter shapes differ slightly between generations — especially reasoning/effort params on the o-series replacements.
• Fine-tunes: triage now. You can't rebuild one tonight. Decide: rebuild on the new base this week, or fall back to the base model temporarily.
• Image pipelines: test gpt-image-2.5 output. Image model swaps change output style — eyeball a few before committing.

What we'd actually do

Found gpt-3.5-turbo in one place? Swap to gpt-5.6-terra, smoke-test, done — 20 minutes, the most common case.

Found a pinned gpt-4o-2024-05-13? Swap to gpt-5.6-sol, but replay production prompts and diff the outputs. Pinned snapshots were pinned for a reason — budget an hour.

Found a fine-tune? Point the call at the base replacement as a stopgap tonight and schedule the rebuild. The stopgap will be dumber than your fine-tune — tell whoever owns that feature.

And write the model ID into one place this time — a config constant, not twelve string literals across four services. Future-you, staring at the Jan 14 4o shutdown, will be grateful.

The bigger pattern

This isn't a one-off. OpenAI's calendar keeps going: evals read-only Oct 31, the prompts API and Agent Builder Nov 30, older image models Dec 1, first-gen GPT-5 snapshots Dec 11, the rest of GPT-4o Jan 14. Model IDs are now a maintenance surface, like dependency versions. Treat a deprecation email like a CVE notice — because tomorrow, for twelve model IDs, it is one.

FAQ