Skip to main content

OpenAI lists GPT-4o API models for January 2027 removal

14 OCTOBER 2026·2 MIN READ·3 SOURCES

OpenAI’s API deprecations page lists several GPT-4o and GPT-4.1 models for removal on January 14, 2027, according to AICoder. The list names GPT-6 Luna as a recommended replacement, while OpenAI’s migration guidance advises testing it on representative workloads before moving production traffic.

OpenAI lists GPT-4o API models for January 2027 removal

Key takeaways · 4

  • 01

    The listed models are scheduled to leave the API on January 14, 2027.

  • 02

    The retirement list includes two GPT-4o snapshots, GPT-4o mini, GPT-4.1 and GPT-4.1 mini.

  • 03

    AICoder reports that new fine-tuning jobs for existing customers will no longer be available from January 6, 2027.

  • 04

    OpenAI recommends testing GPT-6 Luna on representative workloads before moving production traffic.

Which API models are listed

AICoder says OpenAI’s deprecations page, dated October 14, 2026, lists models for removal from the API.[1] The GPT-4o entries are snapshots gpt-4o-2024-08-06 and gpt-4o-2024-11-20, alongside gpt-4o-mini-2024-07-18.[1] The list also includes GPT-4.1 and GPT-4.1 mini, both identified as 2025-04-14 snapshots.[1] AICoder gives January 14, 2027 as the removal date for the listed models.[1]

Fine-tuning has an earlier cutoff

AICoder reports that fine-tuned versions of the retiring models are scheduled to shut down on the same day as their base models: January 14, 2027.[1] It also says existing customers will lose the ability to start new fine-tuning jobs on January 6, eight days earlier.[1] That earlier cutoff makes the fine-tuning date a separate planning milestone from the API retirement itself. Teams that depend on fine-tuned versions can use the listed dates to plan evaluations and migration work; the evidence does not specify a separate transition path for those versions.

Replacement, price and quality

AICoder identifies gpt-6-luna as the recommended replacement for the retiring models.[1] Its listed standard API prices are $0.10 per million input tokens and $0.50 per million output tokens, compared with GPT-4o’s $2.50 and $10 respectively.[1] OpenAI’s Help Center says GPT-6 Luna has a lower per-token price than the GPT-4.x models covered by its migration guidance.[2] However, AICoder says OpenAI publishes no quality comparison between GPT-6 Luna and GPT-4o.[1] Lower listed prices alone therefore do not establish how a particular workload will perform.

Test before moving production traffic

OpenAI’s migration guidance recommends evaluating GPT-6 Luna on representative workloads before moving production traffic.[2] It says most GPT-4.x use cases are expected to migrate with the stated setting.[2] The API timeline is distinct from an earlier ChatGPT change: OpenAI’s Help Center says GPT-4o and additional models were deprecated in ChatGPT on February 13, 2026, while remaining available through the API.[3] ChatGPT Business, Enterprise and Edu customers retained GPT-4o access within Custom GPTs until April 3, 2026.[3] This announcement concerns the API retirement schedule.

The dates give teams using these API models a defined window to test replacements and plan changes to fine-tuning workflows. Lower listed token prices may inform cost planning, but representative workload testing is important because the evidence provides no published quality comparison between GPT-6 Luna and GPT-4o.

Why it matters
Story quiz

Test yourself on this story — 3 questions.

Create a free account to take the quiz, earn XP, and get a daily session built for your industry.

Take the quiz

How this developed

  1. 14 October 2026

    OpenAI lists GPT-4o API models for January 2027 removal

Sources

AI fluency, one session a day, built for your work.