OpenAI said on its API deprecations page that on 14 October 2026 it notified developers using certain GPT-4 series models that they will be removed from the API on 14 January 2027. The models are GPT-4o (two snapshots), GPT-4o mini, GPT-4.1 and GPT-4.1 mini, and the recommended replacement for all of them is gpt-6-luna.
OpenAI’s pricing page lists gpt-6-luna at $0.10 per million input tokens and $0.50 per million output tokens on its Standard tier. The page also says fine-tuned versions of the retired models follow the same shutdown schedule as their base models.
Business Pill · SWITCHING COSTS
A one-minute explainer of switching costs: what it takes for a customer to move from one supplier to another. It teaches the general idea only and says nothing about any organisation in this story.
The key insight: As we read it, the date that matters for most customers is the fine-tuning one: OpenAI’s page says fine-tuned versions follow the same 14 January 2027 shutdown as their base models, and its separate fine-tuning section says active customers can no longer create new fine-tuning jobs on 6 January 2027.
The 14 October entry lists five model snapshots with a shutdown date of 14 January 2027: gpt-4o-2024-08-06 and gpt-4o-2024-11-20, gpt-4o-mini-2024-07-18, gpt-4.1-2025-04-14 and gpt-4.1-mini-2025-04-14. Each row names gpt-6-luna as the recommended replacement.
The same page already lists an earlier set of shutdowns on 23 October 2026, from an entry dated 22 April 2026. That set includes gpt-4o-2024-05-13, gpt-4.1-nano, gpt-4-turbo and several o-series snapshots.
OpenAI’s separate GPT-6 Luna model page describes it as the company’s “most efficient model for focused, high-volume tasks.”
Notice before shutdown, in months: the 14 October 2026 GPT-4 series entry (3) and the 1 October entry (6) on OpenAI’s deprecations page, next to the minimum notice periods the page states. Months for the two entries are our arithmetic from the notice and shutdown dates.
What Happens to Fine-Tuned Models
The page says “Fine-tuned versions of these models follow the same shutdown schedule as their underlying base models.” In other words, a customer’s fine-tuned GPT-4o or GPT-4.1 model stops working on 14 January 2027 along with the base model.
A separate section on self-serve fine-tuning, dated 7 May 2026, says active existing customers will no longer be able to create new fine-tuning jobs on 6 January 2027. It adds that inference on fine-tuned models will be disabled only when the underlying base model is deprecated.
By our reading of those two entries, a team with a fine-tuned GPT-4o model has until 6 January 2027 to start any new fine-tuning job, and its existing fine-tuned model runs until 14 January 2027.
The Notice Period
OpenAI’s page sets out its own minimum notice periods “unless safety or compliance concerns require a faster timeline”: “Generally available models: At least 6 months”, and “Specialized variants of generally available models: At least 3 months.”
By our arithmetic, 14 October 2026 to 14 January 2027 is three months. The 14 October entry does not say which notice category applies to these models or give a reason for the timeline.
For comparison, the entry directly below it, dated 1 October 2026, retires gpt-5.3-codex, gpt-5.4-nano and gpt-5.1 on 1 April 2027 and says that comes “with six months’ notice.” Earlier this month, as we reported on 2 October, OpenAIgave gpt-5.4-cyber 20 days’ notice.
The page also says that in some cases developers may be able to provision dedicated capacity for continued access after a model’s shutdown date, through OpenAI’s sales team.
What the Replacement Costs
On the Standard tier, short context, OpenAI’s pricing page lists gpt-6-luna at $0.10 per million input tokens, $0.01 for cached input, $0.125 for cache writes and $0.50 per million output tokens. We covered GPT-6 Sol and Luna pricing when the models launched in September.
The pricing page we read does not list current prices for the GPT-4o and GPT-4.1 snapshots being retired, so we make no price comparison here. Google’s Gemini 4 Argon, by contrast, launched at an introductory $2 per million input tokens that rises after 30 November.
The Structural Read
As we read it, the notice period is the open question: the page states at least 6 months for generally available models, and the 14 October entry gives three months by our arithmetic without saying which category applies.
As we read it, every retired snapshot points to the same model: the 14 October entry names gpt-6-luna for GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini alike, and OpenAI’s Luna model page calls it its most efficient model for focused, high-volume tasks.
As we read it, the page also offers an exit valve for some customers: it says developers may be able to provision dedicated capacity for continued access after a shutdown date.
OpenAI API deprecations page, 14 October 2026 entry
“Fine-tuned versions of these models follow the same shutdown schedule as their underlying base models.”
Three Implications
API DEVELOPERS The page lists 14 January 2027 as the shutdown date for five GPT-4o and GPT-4.1 snapshots, with gpt-6-luna as the replacement.
TEAMS WITH FINE-TUNES The page says fine-tuned versions shut down with their base models, and new fine-tuning jobs stop for active customers on 6 January 2027.
FINANCE TEAMSOpenAI’s pricing page lists gpt-6-luna at $0.10 per million input tokens and $0.50 per million output tokens on the Standard tier.
The framework puts it this way: “The new job of finance is to manage the ratio between the two — work produced per token burned.”
As we read it, a forced migration is where that ratio gets re-measured: OpenAI’s page moves GPT-4o and GPT-4.1 customers to gpt-6-luna on a fixed date, and each team has to re-test what its workload produces per token on the new model.
gpt-6-luna, OpenAI’s recommended replacement for the retired GPT-4o and GPT-4.1 snapshots, as listed on its API pricing page (Standard tier, short context, per million tokens), read 14 October 2026.
What Is Not Established
The deprecations page does not say how many developers or requests still use the retired models, whether a blog post will follow, or why the notice is three months. It does not say how gpt-6-luna compares with GPT-4o on any customer’s workload.
We did not see the email OpenAI says it sent to affected developers; this piece rests on the public deprecations, pricing and model pages.
Business Engineer Framework
The CFO’s Guide to the Token Economy
A Business Engineer framework on why the unit cost of intelligence is falling while the cost of deploying it rises.
OpenAI will remove GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini from its API on 14 January 2027, with fine-tuned versions following the same schedule, according to its deprecations page entry of 14 October 2026. It recommends gpt-6-luna, listed at $0.10 per million input tokens and $0.50 per million output tokens.
A note on sourcing. On 14 October 2026 we read OpenAI’s API deprecations page, its API pricing page and its GPT-6 Luna model page in full. Dates, model names and prices are OpenAI’s own; the notice-period months are our arithmetic. We did not see the email OpenAI says it sent to developers. Nothing here is investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur, HackerNoon and Search Engine People. Find him on LinkedIn and Substack. See how we source.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
