🔍 Read the full analysis: How The GPT-4o, GPT-4.1, And GPT-5-nano API Changes Affect Costs 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 deprecation schedule, as reported by ThorstenMeyerAI.com on Nov. 1, 2026, says GPT-5, GPT-5 mini and GPT-5-nano leave the API on Dec. 11, followed by listed GPT-4o and GPT-4.1 snapshots on Jan. 14, 2027. The proposed replacements could lower costs for some users and raise them for others; reasoning-token billing and API behavior changes also need testing.
OpenAI’s API deprecation schedule puts GPT-5, GPT-5 mini and GPT-5-nano on a Dec. 11, 2026 retirement date, with specified GPT-4o and GPT-4.1 snapshots scheduled to follow on Jan. 14, 2027, according to a review by ThorstenMeyerAI.com. The listed replacements change token prices unevenly, meaning developers may see lower bills on some workloads and higher costs on others while also adapting to new reasoning-model behavior.
The report says OpenAI’s deprecations page listed the dates and replacement mappings as of Nov. 1, 2026. On Dec. 11, gpt-5, gpt-5-mini and gpt-5-nano are scheduled to be replaced by gpt-5.6-sol, gpt-5.6-terra and gpt-5.6-luna, respectively. On Jan. 14, the listed GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini snapshots are scheduled to move to gpt-6-luna.
Using short-context standard prices per million tokens, the report lists GPT-4o at $2.50 per million input tokens and $10 per million output tokens, compared with $0.10 and $0.50 for GPT-6 Luna. GPT-4.1 is listed at $2 and $8, while GPT-4.1 mini is $0.40 and $1.60; both map to the same replacement prices. GPT-4o mini, at $0.15 input and $0.60 output, is also listed as more expensive than the replacement’s $0.10 and $0.50 rates.
The direction reverses for some GPT-5 users. The report lists GPT-5-nano at $0.05 input and $0.40 output per million tokens, versus $0.20 and $1.20 for GPT-5.6 Luna. GPT-5 mini is listed at $0.25 and $2, versus $2 and $12 for GPT-5.6 Terra. Those figures are per-token rates, not total bills: actual spending depends on input and output volumes, caching, and workload behavior. The source notes that its comparisons use short-context prices for prompts up to 272,000 tokens.
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
How Replacement Rates Change Bills
The price shifts matter because model names are often embedded in production code, tests and automated workflows. A retirement can cause requests to fail if an application continues calling an unavailable model; changing the model ID is only part of the work. Teams also need to test answer quality, latency, output length and downstream formatting before moving live traffic.
For GPT-4o and GPT-4.1 users, the listed input and output rates for GPT-6 Luna are substantially lower than the retired models’ rates. But a lower price per token does not establish that a workload’s total cost will fall: the replacement may use a different number of tokens, and the report says the replacements are reasoning models. The source says hidden reasoning tokens are billed as output, so teams should measure actual usage rather than budget from headline rates alone.
For GPT-5-nano and GPT-5 mini users, the listed replacement rates are higher. The exact budget effect depends on traffic and token mix, but the prices point to a need for workload-specific estimates ahead of the deadline. Fine-tuned versions are also affected, the report says, and it states that the listed replacements do not support fine-tuning.
Products Worth Considering
LG UltraGear 32GX850A-B 32-Inch 4K UHD Glossy OLED Gaming Monitor
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
Deadlines Across OpenAI’s API
The report says some models had already been shut down on Oct. 23, 2026, including gpt-4.1-nano, the original GPT-4o snapshot, gpt-4-turbo, gpt-3.5-turbo, o1 and o4-mini. Applications still calling those IDs may already be failing, according to the source. Its account is based on OpenAI’s deprecations page, which it says it checked on Nov. 1.
The same report distinguishes these API changes from ChatGPT availability: it says GPT-4o, GPT-4.1 and GPT-4.1 mini were removed from ChatGPT on Feb. 13, 2026. The current dates concern API access and therefore affect applications and services built on the API. The report also describes a separate Model Spec revision dated Oct. 29, 2026, which removes much of the detailed style guidance, including a default for Markdown formatting. That creates another reason to test applications that rely on a specific response format rather than assuming it will persist.
“The replacements think before they answer.”
— ThorstenMeyerAI.com
Questions Before the Migration
The source reports the listed retirement dates and rates but does not establish the cost for any particular customer’s workload. That requires measuring token use and testing the replacement against the application’s real prompts. It is also unclear from the report why the Jan. 14 date follows a notice issued on Oct. 14, given its account that OpenAI’s page promises at least six months’ notice for generally available models unless safety or compliance concerns require a faster schedule. The report says the notice does not explain which rule applies.
Behavioral compatibility is another open issue. The source warns that reasoning tokens are billed as output, certain sampling parameters may need to be removed, and a tight output limit can produce an empty answer that is still billed. It does not provide independent workload tests or a quantified estimate of how often those outcomes occur. Developers should verify these points against current OpenAI documentation and their own API responses before making cost or reliability assumptions.
Audit Models and Test Replacements
Teams using affected models should first search code, configuration, test fixtures and automation for exact model IDs, then compare those IDs with OpenAI’s current deprecation notices. The report says OpenAI emails affected customers and documents deprecations on its website. It also says customers with contracts or committed spend may want to ask their account teams about the notice period and whether dedicated capacity is available for continued access.
Before each deadline, developers can run representative requests on the replacement models and record input, output and cached-token use, response quality, latency, formatting and failure cases. They should check whether their integration passes parameters supported by the old model but not the replacement, and review output limits for reasoning workloads. These steps are particularly relevant before the Dec. 11 and Jan. 14 changes; the source also lists a Nov. 30 shutdown for OpenAI’s Evals platform, Agent Builder and reusable prompt objects, which could affect some testing workflows.
Key Questions
When are GPT-4o and GPT-4.1 API models scheduled to retire?
ThorstenMeyerAI.com reports that the listed GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini snapshots are scheduled to retire on Jan. 14, 2027, with gpt-6-luna listed as their replacement.
Will the changes make API use cheaper?
Not for every model or workload. The report lists lower per-million-token rates for GPT-6 Luna than for GPT-4o and GPT-4.1, but higher rates for the GPT-5-nano and GPT-5 mini replacements. Total costs depend on actual token use and workload behavior.
What happens to GPT-5-nano?
The report says gpt-5-nano is scheduled to leave the API on Dec. 11, 2026 and maps to gpt-5.6-luna. Its listed input price rises from $0.05 to $0.20 per million tokens, and its output price rises from $0.40 to $1.20.
Could changing the model affect application behavior?
Yes. The source describes the replacements as reasoning models and flags hidden reasoning tokens billed as output, parameter changes and possible empty billed responses when output limits are too tight. Developers should test their prompts and integration settings before migration.
Are fine-tuned versions included in the retirements?
According to the report, fine-tuned versions follow their base models’ shutdown schedules. It also says the listed replacements do not support fine-tuning, so affected teams should check current OpenAI documentation for migration options.
Source: ThorstenMeyerAI.com
