🔍 Read the full analysis: OpenAI Is Switching Off GPT-4o, GPT-4.1 And GPT-5-nano In Its API. What Breaks When, What It Costs, And How To Migrate 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 schedule says several API models, including GPT-5-nano, GPT-4o and GPT-4.1 snapshots, will be retired in stages, with dates from October 2026 through January 2027. The source lists replacement models and warns that changes in pricing, reasoning behavior and model instructions could affect existing applications; teams should verify the schedule and test migrations against OpenAI’s current documentation.
OpenAI’s API deprecation schedule lists GPT-5-nano and several other models for retirement on December 11, 2026, followed by GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini snapshots on January 14, 2027. The change matters to developers whose applications call those model IDs: after each listed date, affected requests may stop working unless the application has been moved to a supported model. The schedule and replacement details here were summarized by ThorstenMeyerAI.com on November 1.
The report says OpenAI has already retired a group of models on October 23, 2026, including gpt-3.5-turbo, gpt-4-turbo, gpt-4.1-nano, the original gpt-4o-2024-05-13 snapshot, o1 and o4-mini. It says calls using those IDs may fail now. The source attributes the dates to OpenAI’s deprecations page, but developers should check that page directly for the current status of each model.
The next listed date is December 11, 2026. The report says gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-pro, o3 and o3-pro are scheduled to leave the API. It lists gpt-5.6-sol, gpt-5.6-terra and gpt-5.6-luna as replacements for different models. On January 14, 2027, the listed GPT-4o and GPT-4.1 snapshots, along with GPT-4o mini and GPT-4.1 mini snapshots, are scheduled for retirement; the report names gpt-6-luna as their recommended replacement.
The listed IDs are dated model snapshots, while developers may use shorter names in code. According to the report, model pages for the short names already show them as deprecated. It also says fine-tuned versions follow their base models’ schedule, and the proposed replacements do not support fine-tuning. These details could affect applications beyond a simple change to a model string, particularly where a product relies on fine-tuning or a specific model’s response format.
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 Applications
For software teams, the immediate issue is whether production code, scheduled jobs, test fixtures or no-code workflows still call a retiring ID. A hard-coded model name can remain unnoticed until a request fails. The report also describes behavior changes that may require more than swapping IDs: listed replacements are reasoning models, and the source says their hidden reasoning tokens are billed as output. It warns that some sampling parameters may no longer be accepted and that tight output limits can result in an empty response that is still billed.
The economics depend on the model being replaced and the way it is used. The source’s November 1 comparison, using standard short-context prices per million tokens, says GPT-4o costs $2.50 per million input tokens and $10 per million output tokens, while the listed gpt-6-luna replacement costs $0.10 and $0.50. By contrast, it gives gpt-5-nano as $0.05 input and $0.40 output, against $0.20 and $1.20 for gpt-5.6-luna. These are source-reported figures, not a guarantee of future pricing; the report says prices can change.
That means migration can affect availability, output behavior and operating costs at once. A replacement that is cheaper for one workload may be more expensive for another, and a reasoning model may handle formatting or response limits differently. Developers need to test representative requests and review billing rather than assume a change will be transparent to users.
Products Worth Considering

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

LG UltraGear 27-inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDR True Black 400

Samsung 27-Inch Odyssey G5 QHD Gaming Monitor (G50F)
As an affiliate, we earn on qualifying purchases.
Earlier Retirements and Notices
The report distinguishes these API changes from OpenAI’s earlier ChatGPT changes. It says GPT-4o, GPT-4.1 and GPT-4.1 mini were removed from ChatGPT on February 13, 2026; the schedule discussed here concerns API access and therefore applications built by developers, not just models selected directly in ChatGPT.
According to the report, OpenAI’s deprecations page says generally available models receive at least six months’ notice unless safety or compliance concerns call for a faster timetable. The report notes that the October 14 notice for the January 14 retirement is three months, but says the entry does not explain which rule applies. It also says OpenAI may offer some customers dedicated capacity for continued access after a shutdown. Teams with contractual commitments or committed spend may want to ask OpenAI about their specific circumstances.
A separate policy change is reported for the same period: a revised Model Spec dated October 29 removes much of the detailed style guidance, including a default preference for Markdown formatting, and directs models to interpret instructions by intent. The source says OpenAI’s own document acknowledges that models “do not yet fully reflect” the Spec. The change is relevant to applications that depend on consistent formatting, but the report does not establish that it changes the retirement schedule or prove how any particular API model will respond.
Open Questions on Timing and Behavior
The schedule and prices in this account were checked by ThorstenMeyerAI.com on November 1, 2026; they may change. The source does not include the full deprecation notices or independent confirmation of every model’s current status, so developers should verify IDs and dates in OpenAI’s live documentation before making changes.
It is also unclear why the reported notice for the January retirements is shorter than the six-month period cited for generally available models. The report does not identify which safety or compliance exception, if any, applies. Nor does it provide workload-specific migration results, so the listed replacement prices and model names alone cannot establish how an application’s quality, latency or total bill will change.
The source says changes to the Model Spec may make explicit formatting instructions more important, but it does not show how particular prompts will behave after the revision. Teams should treat behavior changes as a testing risk, not as a confirmed outcome for every application.
Audit Models Before Each Deadline
Developers should inventory model IDs across application code, configuration, test data and automation tools, then compare each one with OpenAI’s current deprecation page. The source identifies December 11, 2026, and January 14, 2027, as upcoming API dates in its schedule. It also reports that OpenAI’s Evals platform, Agent Builder and reusable prompt objects are due to shut down on November 30, 2026, with existing evals made read-only on October 31; teams using those tools should check the applicable notices as well.
Before migrating, teams can run representative prompts against the replacement, check structured and formatted outputs, test limits and sampling settings, and compare token usage and cost. They should also confirm whether fine-tuning or other model-specific features are available on the proposed replacement. OpenAI’s live notices and account guidance are the next sources to check, particularly for customers asking about dedicated capacity or the shorter notice period.
Key Questions
When is GPT-5-nano scheduled to leave the API?
The source report lists December 11, 2026, for gpt-5-nano and names gpt-5.6-luna as its replacement. Verify the date and model status in OpenAI’s current deprecation documentation.
When are GPT-4o and GPT-4.1 API snapshots scheduled to retire?
The report lists several GPT-4o and GPT-4.1 snapshots, as well as GPT-4o mini and GPT-4.1 mini snapshots, for January 14, 2027. It names gpt-6-luna as the recommended replacement.
Will changing the model ID be enough?
Not necessarily. The report says replacements are reasoning models and may differ in billing, supported sampling parameters and response behavior. Applications should be tested with representative inputs and output requirements before migration.
Are the replacement models always cheaper?
No. The source’s price comparison gives lower listed token rates for gpt-6-luna than for GPT-4o, but higher listed rates for gpt-5.6-luna than for GPT-5-nano. Prices are time-sensitive, and actual costs depend on usage and the current pricing schedule.
Source: ThorstenMeyerAI.com
