🔍 Read the full analysis: OpenAI’s API Model Switch-Off: Costs, Timelines, And Migration Steps on ThorstenMeyerAI.com
Get smart everyday buys 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 API deprecation schedule sets shutdowns for several models on Dec. 11, 2026, and additional GPT-4o and GPT-4.1 snapshots on Jan. 14, 2027. The source lists replacement models and prices, but warns that reasoning behavior, request settings and output limits may affect migrations; teams should test their applications before switching.
OpenAI has scheduled API shutdowns for GPT-5, GPT-5 mini, GPT-5 nano, o3 and related models on Dec. 11, 2026, followed by listed GPT-4o and GPT-4.1 snapshots on Jan. 14, 2027. The dates affect software and services using those API model IDs, not just people chatting with OpenAI directly, and developers need to test a replacement before affected requests stop working.
According to the deprecation schedule reviewed by ThorstenMeyerAI.com on Nov. 1, models scheduled to leave on Dec. 11 include gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-pro, o3 and o3-pro. The source lists gpt-5.6-sol as the replacement for GPT-5 and o3, gpt-5.6-terra for GPT-5 mini, and gpt-5.6-luna for GPT-5 nano. It lists gpt-5.6-sol with pro reasoning mode for GPT-5 Pro and o3 Pro.
On Jan. 14, 2027, the schedule names dated snapshots of GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini. The article says OpenAI recommends gpt-6-luna for those models. It also reports that a separate group—including gpt-3.5-turbo, gpt-4-turbo, o1 and o4-mini—was already shut down on Oct. 23, so any applications still calling those IDs may already be failing.
The source says fine-tuned versions follow their base models’ shutdown dates and that the listed replacements do not support fine-tuning. It also flags a Nov. 30 shutdown for OpenAI’s Evals platform, Agent Builder and reusable prompt objects, with existing evals read-only from Oct. 31. These changes could affect migration work as well as production calls.
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 Behavior Risks
A model ID change is not necessarily a drop-in substitution. The source describes the listed replacements as reasoning models, which may handle instructions and generate output differently from the systems they replace. It says hidden reasoning tokens are billed as output, some sampling parameters must be removed, and a restrictive output-token limit can result in an empty answer that is still billed. Those are operational and cost risks for teams that change a model name without testing representative requests.
Price effects vary by model. The source’s comparison of short-context API rates, checked Nov. 1, says GPT-4o’s listed replacement, gpt-6-luna, costs $0.10 per million input tokens and $0.50 per million output tokens, compared with $2.50 and $10 for GPT-4o. But it gives different results for some GPT-5 users: gpt-5-nano to gpt-5.6-luna rises from $0.05 to $0.20 per million input tokens and from $0.40 to $1.20 per million output tokens. GPT-5 mini’s listed move to gpt-5.6-terra is described as eight times the input price and six times the output price.
These are per-token prices, not a forecast of total bills. Actual spending depends on usage, prompt and output lengths, caching and the replacement model’s behavior. The source says its figures apply to short-context prices for prompts up to 272,000 tokens and cautions that pricing can change. Organizations should recalculate against their own workloads rather than assume a migration will lower costs.
Products Worth Considering
Apple Mac mini Desktop Computer with M4 chip, 10‑core CPU and GPU, 24GB Unified Memory, 512GB SSD, Gigabit Ethernet
Dell XPS 13 9350 Copilot+ Laptop, 13.4-inch 120Hz Display, Intel Core Ultra 7 256V, 16GB RAM, 512GB SSD, Ubuntu Linux 24.04 LTS
Apple MacBook Air 13-inch with M5 chip, 16GB memory, 1TB SSD, and 12MP Center Stage camera (Midnight)
As an affiliate, we earn on qualifying purchases.
The API Schedule Behind the Deadlines
The schedule in the source is attributed to OpenAI’s deprecations page, which the article says it checked on Nov. 1, 2026. The article reports that OpenAI emails affected customers and documents deprecations on that page. It also notes that the model pages for the short names gpt-4o, gpt-4o-mini, gpt-4.1 and gpt-4.1-mini already display a “Deprecated” status, while the Jan. 14 list names dated snapshots. The source advises treating the short names as part of that deadline; developers should verify the status of the exact IDs in their own applications.
The article distinguishes these API changes from an earlier ChatGPT product change: it says GPT-4o, GPT-4.1 and GPT-4.1 mini were removed from ChatGPT on Feb. 13, 2026. The current schedule is about API access, which supports third-party apps, internal tools and automated workflows. It also reports that a revised Model Spec dated Oct. 29 removes much of the prior detailed style guidance, including a default to format responses in Markdown. Applications that rely on a particular output format should state that requirement explicitly and validate the response.
The source says OpenAI’s deprecation page promises at least six months’ notice for generally available models, unless safety or compliance concerns call for a faster timetable. It observes that the Jan. 14 notice was dated Oct. 14, three months before the scheduled shutdown, but says the entry does not explain which policy applies. OpenAI is also reported to say some customers may be able to purchase dedicated capacity for continued access after a shutdown.
Open Questions on Access and Pricing
Several details remain unresolved in the source material. It does not explain why the Jan. 14 shutdown notice appears to provide three months rather than the stated six-month minimum, or whether OpenAI considers an exception applicable. Customers with contracts or committed spend are advised by the source to ask their account teams about their specific arrangements.
The source provides a pricing snapshot rather than a guarantee of future rates, and does not establish how much a given customer’s total bill will change after migration. It also does not provide a complete compatibility matrix for every API feature or application. The exact effects of the Model Spec revision on individual responses are not quantified; the article says OpenAI acknowledges that model behavior does not yet fully reflect the Spec.
OpenAI’s schedule, replacement recommendations and any eligibility for dedicated capacity are reported here through the supplied source material. The source does not include direct responses from OpenAI addressing the notice-period question or confirming how each customer’s workloads will behave after migration.
Test Before Each Shutdown Date
Teams should first inventory model IDs across code, configuration files, tests, no-code automations and production services, then compare each ID with OpenAI’s current deprecation page. The source’s dates make the next checkpoints Nov. 30 for platform tools, Dec. 11 for the listed GPT-5 and o3 models, and Jan. 14 for the listed GPT-4 snapshots. Because schedules and prices can change, developers should recheck the official listings before acting.
For each affected application, run realistic requests against the recommended replacement in a test environment. Compare output quality, formatting, latency and token use; update parameters the new model does not accept; and review output-token limits to reduce the risk of empty responses. Recalculate costs using current input, output and cached-token rates and the application’s own traffic. If fine-tuning, contracted access or a need to retain an older model is involved, ask OpenAI about available options well before the relevant deadline.
Key Questions
Which OpenAI API models are scheduled to shut down on Dec. 11, 2026?
The source lists gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-pro, o3 and o3-pro for Dec. 11. Check OpenAI’s current deprecation page for the precise IDs and any schedule changes.
When are the listed GPT-4o and GPT-4.1 API snapshots scheduled to stop working?
The source says the listed GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini snapshots are scheduled for shutdown on Jan. 14, 2027. It says OpenAI recommends gpt-6-luna as their replacement.
Will switching to a replacement model automatically reduce API costs?
No. The source reports lower listed rates for some migrations and higher rates for others. For example, it says GPT-5 nano’s move to gpt-5.6-luna raises the listed input and output prices. Total spending also depends on token usage and current rates.
What should developers test before migrating?
Test representative prompts and outputs, required formats, supported request parameters, output limits, latency and token costs. The source warns that reasoning behavior and billing can differ, and that some fine-tuned models may not have a listed replacement that supports fine-tuning.
Source: ThorstenMeyerAI.com
