🔍 Read the full analysis: Migrating Your OpenAI API Apps Before GPT-4o, GPT-4.1, And GPT-5-nano Switch Off 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 API deprecation schedule lists model shutdowns on Oct. 23 and Dec. 11, 2026, and Jan. 14, 2027. The listed replacements may differ in cost and behavior, and the source says some affected models and tools have already become unavailable. Developers should check every model ID in their apps, test replacements and confirm current details against OpenAI’s documentation.
OpenAI’s API deprecation schedule says several model IDs have already stopped working, with further shutdowns scheduled for Dec. 11, 2026, and Jan. 14, 2027. The changes affect applications that call the listed models directly; developers may need to update model settings and test replacement models before their current API calls fail or produce different results.
According to the deprecation schedule as summarized by ThorstenMeyerAI.com on Nov. 1, models including GPT-5, GPT-5 mini, GPT-5-nano, o3 and o3-pro are scheduled to leave the API on Dec. 11. The article lists GPT-5.6 Sol, GPT-5.6 Terra and GPT-5.6 Luna among their replacements. It says the recommended replacement for the GPT-4o and GPT-4.1 models scheduled for Jan. 14 is GPT-6 Luna.
The schedule’s Oct. 23 shutdown date covers a range of older IDs, including GPT-4.1-nano, GPT-4 Turbo, GPT-3.5 Turbo, o1 and o4-mini. The source says calls that still use those IDs may already be failing. It also warns that developers should check dated model snapshots as well as familiar short names, since the listed dates apply to specific model IDs and the article says the related model pages mark several short names as deprecated.
Migration can change bills as well as availability. The source’s comparison of listed standard prices says GPT-4o to GPT-6 Luna would reduce input pricing from $2.50 to $0.10 per million tokens and output pricing from $10 to $0.50. By contrast, it says the move from GPT-5-nano to GPT-5.6 Luna raises the listed input price from $0.05 to $0.20 per million tokens and output from $0.40 to $1.20. These are source-reported prices checked Nov. 1, for short-context use, and may change.
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 Can Change Costs and Outputs
For developers, a model shutdown is not just a change in a configuration label. If an application continues sending requests to a retired ID, the service may stop working. If it is switched to a replacement without testing, differences in response format, latency, reasoning behavior or cost could affect users and downstream systems.
The source describes the listed replacements as reasoning models and says hidden reasoning tokens are billed as output. It also warns that some sampling parameters must be removed and that a restrictive output-token limit can result in an empty answer that is still billed. Those are claims in the source article; developers should verify model-specific behavior in current OpenAI documentation and test against their own workloads.
Pricing changes may cut expenses for some migrations and increase them for others. Teams with high request volumes should calculate costs using their actual input and output patterns, rather than treating a per-million-token rate as a complete forecast. The model change may also require updates to prompts, validation rules and monitoring.
Products Worth Considering
Dell UltraSharp 27 Thunderbolt Hub Monitor (U2726DER)
Dell P2725H 27-Inch Full HD Monitor
Dell UltraSharp 32 U3226QER 4K Thunderbolt Hub Monitor
As an affiliate, we earn on qualifying purchases.
A Separate API Change From ChatGPT
The source distinguishes these API retirements from changes to the ChatGPT product. It says OpenAI removed GPT-4o, GPT-4.1 and GPT-4.1 mini from ChatGPT on Feb. 13, 2026, while the schedule discussed here concerns API access used by apps and bots. A change to ChatGPT availability does not, by itself, describe whether an API model remains available.
There are other dates to track. The source says OpenAI’s Evals platform, Agent Builder and reusable prompt objects are scheduled to shut down Nov. 30, 2026, with existing evals read-only since Oct. 31. It also lists further retirements later in 2027, including some text-to-speech, audio, transcription and newer model IDs. These dates and products are reported from the source’s Nov. 1 review of OpenAI’s deprecation page, and should be rechecked as schedules can be updated.
The article also notes a potential notice-period question: it says OpenAI’s page promises at least six months’ notice for generally available models, subject to safety or compliance exceptions, while the GPT-4o notice it describes runs from Oct. 14 to Jan. 14. The source says the entry does not explain which provision applies. Developers with contractual commitments should ask OpenAI for clarification rather than infer the terms from the dates alone.
“The replacements think before they answer.”
— Thorsten Meyer, ThorstenMeyerAI.com
Schedule Details Still Need Checking
The source article reports the dates and replacement mappings from an OpenAI deprecations page checked on Nov. 1, 2026. This account does not independently establish whether OpenAI has since revised any entry. Developers should check the live schedule and model documentation before making changes, particularly for short names, snapshots and fine-tuned models.
The source says fine-tuned models follow their base model’s shutdown schedule and that listed replacements do not support fine-tuning. It does not provide details for every customer’s account, deployment or contract. It also says some customers may be able to purchase dedicated capacity for continued access, but does not specify eligibility, availability or terms.
It is also unclear from the source how much each replacement will change performance for a particular application. Its price comparisons do not predict total bills, and its discussion of reasoning behavior does not establish how every prompt or workload will respond. The appropriate migration path depends on testing the use case.
Audit Model IDs and Run Tests
Teams should start by searching source code, environment settings, test fixtures, workflow tools and no-code integrations for every model ID in use. Record the exact model name, including dated snapshots, and compare it with OpenAI’s current deprecation schedule. Any ID listed as already retired deserves immediate investigation; the Dec. 11 and Jan. 14 dates provide deadlines for the next groups.
Before switching production traffic, test the replacement with representative inputs. Check output quality, formatting, latency, token usage, error handling and total cost, and update requests if the replacement does not accept the same parameters. The source recommends stating required formats explicitly rather than relying on a default style; it says the Oct. 29 Model Spec revision removed much of its detailed style guidance, including a default for Markdown formatting.
Finally, establish a monitoring and rollback plan where available, update documentation, and confirm contract or dedicated-capacity questions with OpenAI. The immediate next milestone in the reported schedule is the Nov. 30 shutdown of Evals, Agent Builder and reusable prompt objects, followed by model retirements on Dec. 11 and Jan. 14. Dates and prices should be verified again before each migration.
Key Questions
When are GPT-5-nano and other GPT-5 API models scheduled to shut down?
The source reports that GPT-5, GPT-5 mini and GPT-5-nano, along with several other models, are scheduled to leave the API on Dec. 11, 2026. Check OpenAI’s current deprecation page for updates.
When are the GPT-4o and GPT-4.1 API models scheduled to shut down?
The source lists remaining GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini snapshots for Jan. 14, 2027. It reports GPT-6 Luna as the recommended replacement for those listed models.
Will the replacement model cost less?
Not in every case. The source’s Nov. 1 price comparison reports lower listed rates for some moves, such as GPT-4o to GPT-6 Luna, but higher rates for GPT-5-nano to GPT-5.6 Luna. Actual spending depends on usage and current prices.
What should developers test before migrating?
Test representative requests for output quality, formatting, latency, token use, errors and cost. Check whether the replacement accepts the same parameters and whether your application relies on a model’s default response style.
Are these shutdowns the same as ChatGPT model removals?
No. The source describes these dates as applying to OpenAI API models, which power third-party applications and services. It separately reports earlier removals from ChatGPT, a different product surface.
Source: ThorstenMeyerAI.com
