All guidesVercel and env vars
On this page

OpenAI change

OpenAI model shutdowns: 2026-10-23, 2026-12-11 and 2027-01-14

By Ohad Krispin, founder of VibeRaven · LinkedIn

OpenAI removed gpt-3.5-turbo, gpt-4, gpt-4-turbo, o1, o3-mini, o4-mini, gpt-4.1-nano and gpt-image-1 from the API on October 23, 2026, removes gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-pro, o3 and o3-pro on December 11, 2026, and removes the dated snapshots gpt-4o-2024-08-06, gpt-4o-2024-11-20, gpt-4o-mini-2024-07-18, gpt-4.1-2025-04-14 and gpt-4.1-mini-2025-04-14 on January 14, 2027. Requests to a removed model fail, so replace each ID with the model in the table below before its date.

Next step in your repo: run npx -y viberaven@1.7.8 changes. It lists each OpenAI model ID from these three dates in your code or env templates as openai-model-shutdowns-2026, with the file, line and date: on a fixture repo on 2026-11-03 it printed "lib/ai.ts:7: uses gpt-4 (shut down 2026-10-23)". Its checks are scoped to Vercel + Supabase apps and read repository files only: advice, not a gate, and not a live database test.

Why it matters

When each group was announced

OpenAI announced the first group on April 22, 2026, the GPT-5 and o3 snapshots on June 11, 2026 and the GPT-4o and GPT-4.1 snapshots on October 14, 2026. Each alias in the table covers its dated snapshot too, such as gpt-4-0613 for gpt-4, and fine-tuned models follow the shutdown date of their base model. Checked against OpenAI's deprecations page on 2026-11-03.

Find the model IDs in your code

Search for each ID as a whole quoted string, so gpt-4 does not also match gpt-4o or gpt-4.1. Model names often sit in an environment variable such as OPENAI_MODEL, so also check the values set in your hosting provider, which the repository does not show.

grep -rnE --exclude-dir=node_modules "['\"](gpt-3\.5-turbo|gpt-4|gpt-4-turbo|gpt-4\.1-nano|o1|o1-pro|o3-mini|o4-mini|gpt-image-1|gpt-5|gpt-5-mini|gpt-5-nano|gpt-5-pro|o3|o3-pro)['\"]|gpt-4o-2024-(05-13|08-06|11-20)|gpt-4o-mini-2024-07-18|gpt-4\.1(-mini)?-2025-04-14" .

The bare gpt-4o and gpt-4o-mini aliases

The January 14, 2027 entry names the dated snapshots gpt-4o-2024-08-06, gpt-4o-2024-11-20 and gpt-4o-mini-2024-07-18, not the bare aliases gpt-4o and gpt-4o-mini, which are not on the deprecations page. Code that pins a dated snapshot needs a change before that date; code that sends the bare alias is not named there.

Other OpenAI dates in the same months

The same page lists more dates that this table leaves out: gpt-image-1-mini, gpt-image-1.5 and chatgpt-image-latest on December 1, 2026; the v1/prompts API, the Evals platform and Agent Builder on November 30, 2026; tts-1 and tts-1-hd on January 6, 2027; and whisper-1 and gpt-4o-transcribe on February 26, 2027. Azure OpenAI deployments and IDs sent through a gateway such as OpenRouter follow their own schedules.

Checklist

Before each shutdown date

  1. 01Search the code, env templates and your hosting provider's environment variables for each ID in the table.
  2. 02Replace each ID with its recommended replacement; for o1-pro, gpt-5-pro and o3-pro, set reasoning.mode to pro as the table says.
  3. 03Run each AI feature once: compare the output, the token use and the cost per request.
  4. 04Deploy before the date, then watch the logs for errors that name a model.

Shutdown table

OpenAI model IDs and their replacements

Copied from OpenAI's deprecations page on 2026-11-03. A date in the past means requests to that model already fail.

Model ID, shutdown date and recommended replacement

OpenAI model shutdowns, read 2026-11-03
Model ID (and snapshot)Shutdown dateRecommended replacement
gpt-3.5-turbo (gpt-3.5-turbo-0125)October 23, 2026gpt-5.6-terra
gpt-4 (gpt-4-0613)October 23, 2026gpt-5.6-sol
gpt-4-1106-previewOctober 23, 2026gpt-5.6-sol
gpt-4-turbo (gpt-4-turbo-2024-04-09)October 23, 2026gpt-5.6-sol
gpt-4.1-nano (gpt-4.1-nano-2025-04-14)October 23, 2026gpt-5.6-luna
gpt-4o-2024-05-13October 23, 2026gpt-5.6-sol
gpt-image-1October 23, 2026gpt-image-2.5-sunburst or gpt-image-2.5-flare
o1 (o1-2024-12-17)October 23, 2026gpt-5.6-sol
o1-pro (o1-pro-2025-03-19)October 23, 2026gpt-5.6-sol (reasoning.mode: pro)
o3-mini (o3-mini-2025-01-31)October 23, 2026gpt-5.6-sol
o4-mini (o4-mini-2025-04-16)October 23, 2026gpt-5.6-terra
gpt-5 (gpt-5-2025-08-07)December 11, 2026gpt-5.6-sol
gpt-5-mini (gpt-5-mini-2025-08-07)December 11, 2026gpt-5.6-terra
gpt-5-nano (gpt-5-nano-2025-08-07)December 11, 2026gpt-5.6-luna
gpt-5-pro (gpt-5-pro-2025-10-06)December 11, 2026gpt-5.6-sol (reasoning.mode: pro)
o3 (o3-2025-04-16)December 11, 2026gpt-5.6-sol
o3-pro (o3-pro-2025-06-10)December 11, 2026gpt-5.6-sol (reasoning.mode: pro)
gpt-4o-2024-08-06, gpt-4o-2024-11-20January 14, 2027gpt-6-luna
gpt-4o-mini-2024-07-18January 14, 2027gpt-6-luna
gpt-4.1-2025-04-14January 14, 2027gpt-6-luna
gpt-4.1-mini-2025-04-14January 14, 2027gpt-6-luna

Related

Sources

Official docs behind this checklist

FAQ

FAQ for builders before launch

Do fine-tuned models shut down too?

Yes. OpenAI says fine-tuned versions of these models follow the same shutdown schedule as their base models.

Is gpt-4o shut down?

The dated snapshots gpt-4o-2024-05-13 (October 23, 2026), gpt-4o-2024-08-06 and gpt-4o-2024-11-20 (January 14, 2027) are on the page. The bare alias gpt-4o is not.