🔍 Read the full analysis: Three Models, One Migration: Preparing For OpenAI’s API Switch-Off on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
OpenAI’s API deprecation schedule puts several model shutdowns on 11 December 2026 and 14 January 2027, while some older model IDs stopped working on 23 October. The listed replacements can change costs and output behavior, so developers need to check model IDs, update integrations and test before their deadlines.
OpenAI has scheduled API shutdowns for multiple model families, with GPT-5, GPT-5 mini, GPT-5 nano, GPT-5 Pro, o3 and o3 Pro set to leave on 11 December 2026, followed by GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini snapshots on 14 January 2027. Some older IDs, including GPT-3.5 Turbo and the original GPT-4o snapshot, were already shut down on 23 October, according to the company’s deprecation schedule as reviewed by Thorsten Meyer on 1 November.
The schedule names gpt-5.6-sol, gpt-5.6-terra, gpt-5.6-luna and gpt-6-luna as replacement options for affected models. OpenAI’s listed mapping sends GPT-5 and o3 to gpt-5.6-sol; GPT-5 mini to gpt-5.6-terra; and GPT-5 nano to gpt-5.6-luna. GPT-5 Pro and o3 Pro are mapped to gpt-5.6-sol with its pro reasoning mode. The GPT-4o and GPT-4.1 families are mapped to gpt-6-luna.
Those substitutions may change operating costs. At the listed short-context rates per million tokens, Meyer reports GPT-4o at $2.50 for input and $10 for output, compared with $0.10 and $0.50 for gpt-6-luna. The direction reverses for some GPT-5 users: GPT-5 nano’s listed replacement costs $0.20 per million input tokens and $1.20 per million output tokens, versus $0.05 and $0.40 for the retiring model. GPT-5 mini’s replacement is listed at $2 and $12, compared with $0.25 and $2. Prices are dated observations, not guarantees of future rates.
The replacements are described as reasoning models. Meyer warns that hidden reasoning tokens count as billed output, some sampling parameters may need to be removed, and a restrictive output-token cap can produce an empty answer that is still billed. The article also notes that OpenAI’s Model Spec, revised on 29 October, removed much of its detailed style guidance, including a default for Markdown formatting. Apps that depend on a particular format may need to specify it in their requests.
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 Costs and Output Changes
This is an operational change for companies whose software, automations or customer-facing services call the affected API model IDs. A retired ID can cause requests to fail; replacing it with a newer model may also alter response style, token use and billing. The change is not limited to large applications: an old model name can remain in a configuration file, test fixture or low-code workflow long after its original owner has moved on.
The price examples show why teams should measure their own workloads rather than assume a newer option will be cheaper. Some GPT-4-series migrations have lower listed token rates, while the listed GPT-5 nano and GPT-5 mini replacements cost more per token. Reasoning tokens and changed output limits can further affect the final bill. These figures do not establish what a particular customer will pay overall; usage, prompt length, caching, output volume and any contract terms matter.
Behavior changes also matter where applications rely on consistent formatting or predictable responses. A migration test should check whether outputs still meet the application’s requirements, not just whether an API call succeeds. That is especially relevant for structured outputs and automations that pass model responses directly to another system.
Products Worth Considering
Samsung 27-inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
Samsung 32-inch Odyssey OLED G8 (G80SH) 4K QD-OLED Gaming Monitor
LG UltraGear 34GX90SA-W 34-inch WQHD Curved OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
The API Deprecation Timeline
The dates in Meyer’s report are drawn from OpenAI’s deprecations page, which he says he checked on 1 November 2026. The page lists several models as already shut down on 23 October, including gpt-3.5-turbo, gpt-4-turbo, gpt-4.1-nano, gpt-4o-2024-05-13, o1, o1-pro, o3-mini, o4-mini and gpt-image-1. If those IDs remain in an active integration, the article says requests using them may fail.
The next two principal dates are 11 December 2026 and 14 January 2027. The report also flags a separate 30 November shutdown affecting OpenAI’s Evals platform, Agent Builder and reusable prompt objects. Evals became read-only on 31 October, according to the source. Other audio, speech and coding-related models have later 2027 dates, including listed shutdowns in January, February and April.
This API schedule is distinct from changes to ChatGPT. Meyer reports that OpenAI removed GPT-4o, GPT-4.1 and GPT-4.1 mini from ChatGPT on 13 February 2026; the current schedule concerns developers and services built on the API. The report says fine-tuned models follow their base model’s shutdown dates and that the listed replacements do not support fine-tuning.
Notice Period and Migration Details
The schedule and pricing described here are based on a 1 November 2026 check and may change. The source says OpenAI’s general policy promises at least six months’ notice for generally available models unless safety or compliance concerns require a faster timetable. It also notes that the 14 October notice for the 14 January shutdown spans three months, but says the entry does not explain which policy condition applies.
It is not clear from the supplied material whether every affected customer has received a direct notice, whether any individual customer can retain access under a contract, or what continued access through dedicated capacity would cost. Meyer says OpenAI indicates that some customers may be able to buy such capacity. The source also does not provide independent test results comparing the replacements’ accuracy, latency or performance on specific applications.
OpenAI’s revised Model Spec describes intended behavior, but the company’s statement that models “do not yet fully reflect” it means developers should not treat the document as a guarantee of uniform outputs. The precise behavior and billing impact of a replacement will depend on the model, settings and workload.
Audit IDs Before Each Deadline
Teams should first search application code, configuration files, test fixtures and no-code workflows for every model ID in use, including dated snapshots and short names. The source says the short names for GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini are marked deprecated and should be treated as part of the January schedule. Any ID already retired should be investigated immediately.
Before moving traffic, developers can test the mapped replacement against representative prompts and outputs, checking formatting, application-specific requirements, token use and costs. They should review sampling settings and output-token limits, since reasoning models may handle those differently, and update any process that depends on OpenAI’s Evals, Agent Builder or reusable prompt objects before the 30 November date. Teams with committed spend or questions about notice and continued access can ask their OpenAI account team; pricing and the deprecation page should be checked again before deployment.
Key Questions
When are the main API shutdown dates?
The main dates are 11 December 2026 for GPT-5, GPT-5 mini, GPT-5 nano, GPT-5 Pro, o3 and o3 Pro, and 14 January 2027 for the listed GPT-4o and GPT-4.1 snapshots. Some other model IDs were listed as shut down on 23 October 2026.
Which models replace the retiring GPT-4o and GPT-4.1 versions?
According to the mapping reported by Thorsten Meyer, the listed replacement for the affected GPT-4o and GPT-4.1 family snapshots is gpt-6-luna. Developers should check OpenAI’s current model and deprecation pages before changing production systems.
Will migration always reduce API costs?
No. The listed rates are lower for some GPT-4-series replacements, but higher for the reported GPT-5 nano and GPT-5 mini replacements. Actual spending will also depend on usage and billed reasoning output, and the prices may change.
What should developers test before switching?
Test the replacement with representative prompts and check response format, application behavior, token use and cost. Review sampling parameters and output limits, and search code, configuration files and no-code tools for model IDs that may not be in active development.
Source: ThorstenMeyerAI.com
