🔍 Read the full analysis: OpenAI’s Model Switch-Off: How To Migrate Before GPT-4o, GPT-4.1, And GPT-5-nano Go on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
OpenAI’s published deprecation schedule lists API model retirements for Dec. 11, 2026, and Jan. 14, 2027; several other IDs stopped working on Oct. 23. The listed replacements can differ in price and behavior, so developers should inventory model IDs, test replacement models and verify current pricing before switching.
OpenAI’s API deprecation schedule sets Dec. 11, 2026, for the retirement of GPT-5, GPT-5 mini, GPT-5-nano and several o3 models, followed by listed GPT-4o and GPT-4.1 snapshots on Jan. 14, 2027. The schedule also includes model IDs that stopped working on Oct. 23, meaning developers may already need to fix affected applications.
According to the deprecation page reviewed by ThorstenMeyerAI.com on Nov. 1, the December group includes gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-pro, o3 and o3-pro. The listed replacements are in OpenAI’s GPT-5.6 family: for example, gpt-5-nano maps to gpt-5.6-luna, while gpt-5-mini maps to gpt-5.6-terra. The source says fine-tuned models are also scheduled to retire with their base models, and that the listed replacements do not support fine-tuning.
On Jan. 14, 2027, the schedule lists dated snapshots of GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini. It recommends gpt-6-luna for those entries. Earlier retirements dated Oct. 23 include gpt-4.1-nano, the original gpt-4o-2024-05-13 snapshot, gpt-4-turbo, gpt-3.5-turbo, o1 and o4-mini. The article says calls to those IDs may fail if they remain in use.
The source’s pricing comparison, based on standard short-context rates per million tokens checked Nov. 1, shows a wide range of changes. It lists GPT-4o at $2.50 per million input tokens and $10 per million output tokens, compared with $0.10 and $0.50 for gpt-6-luna. By contrast, it lists GPT-5-nano at $0.05 input and $0.40 output, compared with $0.20 and $1.20 for gpt-5.6-luna. These are figures reported by the source, not a guarantee of future prices; developers should check OpenAI’s pricing page before budgeting.
Infographic 01 · The shutdown calendar
What stops working, and when
OpenAI's API deprecations, as listed on 1 November 2026. GPT-4o and GPT-4.1 already left ChatGPT on 13 February 2026; this round is about apps built on the API.
- 23 Oct 2026Already offAlready shut down
gpt-3.5-turbo,gpt-4,gpt-4-turbo,gpt-4.1-nano,gpt-4o-2024-05-13,o1,o1-pro,o3-mini,o4-mini,gpt-image-1. Calls to these fail today. - 30 Nov 2026Your testsEvals platform, Agent Builder, reusable prompts
Not models, but tools many teams test with. Existing evals became read-only on 31 Oct. Export them first.
- 11 Dec 2026Notified 11 JunGPT-5 family and o3
gpt-5→gpt-5.6-sol·gpt-5-mini→gpt-5.6-terra·gpt-5-nano→gpt-5.6-luna·o3→gpt-5.6-sol·gpt-5-pro,o3-pro→gpt-5.6-sol(pro mode) - 14 Jan 2027Notified 14 OctGPT-4o, GPT-4o mini, GPT-4.1, GPT-4.1 mini
Snapshots
gpt-4o-2024-08-06,gpt-4o-2024-11-20,gpt-4o-mini-2024-07-18,gpt-4.1-2025-04-14,gpt-4.1-mini-2025-04-14→gpt-6-luna. Fine-tuned versions follow their base model. - Jan to Apr 2027Plan aheadAudio, speech, older GPT-5.x
6 Jan:
tts-1and others · 20 Jan: older audio and realtime models · 26 Feb:whisper-1, GPT-4o transcription · 1 Apr:gpt-5.1,gpt-5.4-nano,gpt-5.3-codex
Sources: OpenAI deprecations page (checked 1 Nov 2026) · OpenAI Help Center, model release notes. Dates as listed by OpenAI; shutdowns can be rescheduled.
Powered by Thorsten Meyer AI
Migration Risks for API Users
For companies using OpenAI through the API, a model retirement can affect production applications, internal tools and automated workflows, not just a model choice in a user interface. A hard-coded identifier in a service, test fixture or no-code automation may go unnoticed until a request fails. The Oct. 23 date makes an inventory urgent for teams that have not checked their model use recently.
Changing the model ID alone may not preserve an application’s behavior or cost. The source describes the listed replacements as reasoning models, which may use hidden reasoning tokens billed as output. It also warns that some sampling parameters must be removed and that a tight output-token limit can result in an empty answer that is still billed. Those differences can affect latency, output format, token use and the application’s handling of responses.
The price examples underline why teams should test each use case rather than assume migration will make every workload cheaper. The source’s listed GPT-4o replacement costs less per token, while its GPT-5-nano replacement costs more on both the input and output rates shown. The effect on a particular bill will depend on actual usage and the current pricing terms.
Products Worth Considering
Samsung Galaxy Tab S10 FE 128GB Wi-Fi Android Tablet, Blue
DJI Osmo Mobile 8 Standard Combo Gimbal Stabilizer for Phone
Amazon Kindle Paperwhite (12th Generation, 2024), 16 GB, 7-inch Display, Jade
As an affiliate, we earn on qualifying purchases.
Published Retirement Schedule
The schedule in the source separates already retired models from upcoming groups. It says OpenAI announced the Oct. 23 retirements in April, the Dec. 11 group on June 11, and the Jan. 14 group on Oct. 14. OpenAI’s deprecations page is the source of those dates, according to the article, and OpenAI says it emails affected customers and documents deprecations there.
The source also reports that ChatGPT’s model removals are a separate matter: OpenAI removed GPT-4o, GPT-4.1 and GPT-4.1 mini from ChatGPT on Feb. 13, 2026. The current schedule discussed here concerns API access and the applications developers have built with it.
A separate platform change is scheduled for Nov. 30, 2026, when OpenAI plans to shut down its Evals platform, Agent Builder and reusable prompt objects at v1/prompts. The source says existing evals became read-only on Oct. 31. That timing may complicate migration testing for users who rely on those tools, so they should check OpenAI’s notices for details relevant to their accounts.
Costs and Behavior Still Need Testing
Several practical details remain unsettled for individual developers. The source says OpenAI’s notice promises at least six months’ notice for generally available models, unless safety or compliance concerns require a faster timeline. It notes that the Oct. 14 notice for the Jan. 14 retirements gives three months, but the notice does not explain which exception or policy applies. Developers with contracts or committed spend may need to ask their OpenAI account team how the schedule affects them.
The source also says some customers may be able to buy dedicated capacity for continued access after a shutdown, but does not specify eligibility, terms or availability. Its reported replacement prices are a snapshot checked Nov. 1, not a fixed future rate. Teams should confirm the applicable price and model availability directly with OpenAI.
Behavior changes are also not fully predictable from model names alone. The source reports that OpenAI’s Oct. 29 Model Spec revision removed much of its detailed style guidance, including a default preference for Markdown, and instructs models to interpret instructions by intent. It does not establish how any particular application’s output will change. Developers need to test their own prompts, formatting requirements and downstream systems.
Audit IDs and Test Replacements
Developers should start by searching source code, configuration files, prompt tools and no-code automations for every model ID they use, including dated snapshots and fine-tuned variants. Compare that list with OpenAI’s current deprecations page and account notices, then prioritize IDs already retired and those scheduled for Dec. 11.
For each affected workflow, test the replacement using representative inputs and the application’s actual output limits. Check whether calls still include sampling parameters the source says some reasoning models do not support; review empty responses, formatting, latency and token use; and confirm that any required output format is stated explicitly in the prompt. Run the results through the product’s normal checks rather than relying on a single successful request.
Finally, record the model, test date and observed price and behavior, and repeat the review before each shutdown date. Teams using OpenAI’s Evals platform, Agent Builder or reusable prompt objects should check the Nov. 30 closure details and arrange an alternative testing process. OpenAI’s notices and pricing page remain the source for account-specific status and current terms.
Key Questions
When are the GPT-5 and GPT-4o API models scheduled to retire?
According to the schedule cited by ThorstenMeyerAI.com, listed GPT-5 and o3 models are scheduled to retire on Dec. 11, 2026. Listed GPT-4o and GPT-4.1 snapshots are scheduled for Jan. 14, 2027.
Which models had already stopped working?
The source reports that several IDs, including gpt-4.1-nano, gpt-4-turbo, gpt-3.5-turbo, o1 and o4-mini, were scheduled to shut down on Oct. 23, 2026. Check OpenAI’s current deprecations page to confirm an ID’s status.
Are the listed replacements always cheaper?
No. The source’s Nov. 1 price comparison lists a lower per-token rate for GPT-4o’s recommended replacement, but higher listed input and output rates for the GPT-5-nano replacement. Prices and actual costs depend on current rates and usage.
What should developers test before migrating?
Test representative requests, response formatting, output limits, supported parameters, empty responses, token use and costs. The source also says the listed replacements do not support fine-tuning, so teams using fine-tuned models should verify their options with OpenAI.
Source: ThorstenMeyerAI.com
