Prepare Your Apps For GPT-4o, GPT-4.1, And GPT-5-nano Leaving The API

🔍 Read the full analysis: Prepare Your Apps For GPT-4o, GPT-4.1, And GPT-5-nano Leaving The API on ThorstenMeyerAI.com

OpenAI plans to retire GPT-5, GPT-5 mini and GPT-5-nano from its API on Dec. 11, 2026, followed by listed GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini snapshots on Jan. 14, 2027, according to the company’s deprecation schedule reviewed by ThorstenMeyerAI.com. The changes affect applications and services that call those API model IDs; some earlier model IDs on the schedule have already stopped working.

The Dec. 11 retirement list also includes GPT-5 Pro, o3 and o3-pro. OpenAI’s listed replacements are GPT-5.6 Sol for GPT-5 and o3, GPT-5.6 Terra for GPT-5 mini, and GPT-5.6 Luna for GPT-5-nano. The article says the pro models are directed to GPT-5.6 Sol with its pro reasoning mode. On Jan. 14, the listed GPT-4o and GPT-4.1 snapshots are directed to GPT-6 Luna.

Several IDs, including GPT-4.1-nano, the original GPT-4o snapshot, GPT-4 Turbo, GPT-3.5 Turbo, o1 and o4-mini, were listed as shut down on Oct. 23, 2026. The source says calls that still use those IDs may already fail. It also reports an Oct. 31 change that made existing OpenAI Evals read-only, ahead of the planned Nov. 30 shutdown of Evals, Agent Builder and reusable prompt objects. These dates and model mappings come from the source’s account of OpenAI’s deprecations page, checked Nov. 1.

The source’s pricing comparison, based on standard short-context prices per million tokens, shows different effects across migrations. GPT-4o is listed 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, GPT-5-nano’s listed replacement, GPT-5.6 Luna, costs $0.20 per million input tokens and $1.20 per million output tokens, compared with $0.05 and $0.40 for GPT-5-nano. GPT-5 mini’s replacement, GPT-5.6 Terra, is listed at $2 and $12, compared with $0.25 and $2. Prices can change; the source advises checking the official pricing page before budgeting.

Migration Can Change Cost and Output

These retirements are not just a matter of replacing one string in a configuration file. Model changes can affect bills, response behavior and application reliability, especially when a service relies on a specific output format or has a fixed token limit. The source describes the recommended replacements as reasoning models. It warns that hidden reasoning tokens are billed as output, that some sampling parameters may need to be removed, and that a tight output limit can result in an empty answer that is still billed.

The pricing examples show why teams should calculate their own usage rather than assume a migration will cost less. The listed GPT-4o replacement is cheaper per token in the source’s comparison, while the GPT-5-nano and GPT-5 mini replacements cost more per token. Actual spending will depend on token volumes, prompt sizes, caching and the model’s output. The comparison is a snapshot of listed prices, not a forecast of a customer’s bill.

Compatibility matters as well as price. A replacement can produce a different structure, tone or answer even when an application request succeeds. That can affect downstream tools, automated workflows and customer-facing responses. Teams that use fine-tuned versions also need to check carefully: the source reports that fine-tuned models follow their base model’s shutdown schedule and that the listed replacements do not support fine-tuning.

Popular right now

Vitamix Propel Series 750 Professional-Grade Blender, 4 Programs, Black

AmazonView Latest Price

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation

AmazonView Latest Price

De'Longhi 15-Bar Pump Espresso Machine with Milk Frother & Temperature Control

AmazonView Latest Price

As an affiliate, we earn on qualifying purchases.

API Changes Follow ChatGPT Removal

This schedule concerns OpenAI’s API, not a new announcement about the ChatGPT interface. ThorstenMeyerAI.com reports that OpenAI removed GPT-4o, GPT-4.1 and GPT-4.1 mini from ChatGPT on Feb. 13, 2026; the API dates apply to developers whose applications call those models. The distinction matters because a change to the consumer product does not by itself establish when an API integration stops working.

The source says the retirement dates cover dated snapshots, while developers may use shorter names such as “gpt-4o” in code. It reports that model pages for the short names already displayed a “Deprecated” status and advises teams to treat those names as affected by the dates attached to their snapshots. The schedule also continues beyond January: the article lists later 2027 dates for text-to-speech, audio, transcription and other models, including GPT-5.1 and GPT-5.4-nano on April 1.

In a separate development, the source reports that OpenAI revised its public Model Spec on Oct. 29, 2026. The updated document removes much of the detailed style guidance, including a default preference for Markdown, and says models should interpret instructions according to their intent. ThorstenMeyerAI.com says OpenAI also acknowledged that its models “do not yet fully reflect” the Spec. For developers, the practical takeaway is to specify required formatting directly rather than rely on an assumed house style.

„The model should read instructions for their intent.“

— ThorstenMeyerAI.com, summarizing the revised Model Spec

Notice Period and Migration Details

OpenAI’s schedule and the replacement mappings are reported by ThorstenMeyerAI.com, which says it checked the deprecations page on Nov. 1, 2026. The source does not establish whether every listed replacement will behave equivalently for a particular application, or what a customer’s eventual usage costs will be. Those depend on the workload and current prices.

The article also points to a notice-period question. It says OpenAI’s page promises at least six months’ notice for generally available models unless safety or compliance concerns call for a shorter timeline. The GPT-4o and GPT-4.1 notice was dated Oct. 14 for a Jan. 14 retirement, a three-month interval. The source says the notice entry does not explain which rule applies. Customers with contractual commitments or committed spend may want to ask their OpenAI account team about their situation; the source reports that some customers may be able to buy dedicated capacity for continued access after a shutdown.

It is also not clear from the supplied material whether all accounts will receive the same communications, whether any dates or mappings will change, or how specific applications will perform after migration. The source recommends checking OpenAI’s current documentation rather than treating its Nov. 1 review as a guarantee of future availability or price.

Audit IDs and Test Replacements

Developers should start by searching code, configuration files, test fixtures, scheduled jobs and no-code workflows for model IDs, including dated snapshots and fine-tuned models. Compare every match against the current OpenAI deprecation schedule, with immediate attention to IDs listed as shut down on Oct. 23 and the next dates of Dec. 11, 2026, and Jan. 14, 2027. Teams should also check whether tests or workflows depend on Evals, Agent Builder or reusable prompt objects before the Nov. 30 date cited in the source.

For each affected integration, record the current model, the proposed replacement, expected input and output volume, and any features or parameters the request uses. Run the replacement against representative inputs and test the parts of the response that matter to the application: formatting, structured data, tool use, refusals, latency and behavior when prompts approach output limits. Recheck the applicable model documentation and prices at the time of migration, and monitor results after deployment. The source’s schedule sets the dates; it does not say that a replacement will be behaviorally identical.

Key Questions

When is GPT-5-nano scheduled to leave the API?

Dec. 11, 2026, according to OpenAI’s schedule as reported by ThorstenMeyerAI.com. The listed replacement is GPT-5.6 Luna.

When are the listed GPT-4o and GPT-4.1 snapshots scheduled to be retired?

The article lists Jan. 14, 2027 for the specified GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini snapshots. It reports GPT-6 Luna as the recommended replacement for those entries.

Have any of the affected models already stopped working?

Yes. The source says several older IDs, including GPT-4.1-nano, the original GPT-4o snapshot, GPT-4 Turbo, GPT-3.5 Turbo, o1 and o4-mini, were shut down on Oct. 23, 2026. Check the current deprecation schedule and your own requests to confirm whether an integration is affected.

Will migration lower API costs?

Not in every case. The source’s price snapshot lists GPT-6 Luna below GPT-4o’s per-token prices, but GPT-5.6 Luna above GPT-5-nano’s and GPT-5.6 Terra above GPT-5 mini’s. Actual costs depend on usage and current pricing.

What should developers test before switching?

Test representative prompts and application outputs, including formatting, structured responses, tool use, parameters, latency and token limits. Also check fine-tuning compatibility, usage-based costs and OpenAI’s latest model documentation before deploying.

Source: ThorstenMeyerAI.com