OpenAI
OpenAI shuts down GPT-4, o1 and o3-mini in the API on Oct. 23
OpenAI ended API access to GPT-3.5 Turbo, GPT-4, o1, o3-mini and GPT Image 1 on Oct. 23, 2026. The text models point to GPT-5.6 models.
OpenAI shut down API access to GPT-3.5 Turbo, GPT-4, GPT-4 Turbo, o1, o3-mini, o4-mini and GPT Image 1 on Oct. 23. The text models point to GPT-5.6 models. GPT Image 1 points to gpt-image-2.5-sunburst or gpt-image-2.5-flare. Developers still calling these models have to switch.
The shutdown dates come from OpenAI’s deprecations page, read Oct. 23. An entry dated April 22, 2026, headed “Legacy GPT model snapshots,” says access ends on the listed dates.
Most Text Models Move to GPT-5.6 Sol
OpenAI lists GPT-5.6 Sol as the replacement for gpt-4-0613, gpt-4-1106-preview, gpt-4-turbo, gpt-4o-2024-05-13, o1, o3-mini and o1-pro. For o1-pro the page adds a setting, reasoning.mode: pro.
GPT-5.6 Terra replaces gpt-3.5-turbo-0125 and o4-mini. GPT-5.6 Luna replaces gpt-4.1-nano. The page also lists the aliases and completions variants of these models.
Fine-Tuned Models Go the Same Day
Fine-tuned versions were removed Oct. 23 along with their bases. ft-gpt-4 points to Sol, and ft-gpt-3.5-turbo, ft-babbage-002, ft-davinci-002 and ft-o4-mini-2025-04-16 point to Terra. ft-gpt-4.1-nano-2025-04-14 points to Luna.
GPT Image 1 Is Gone Too
OpenAI’s public OpenAPI spec, read Oct. 23 on GitHub, says “GPT Image 1 was shut down on October 23, 2026. Use gpt-image-2.5-sunburst or gpt-image-2.5-flare instead.”
A separate deprecations entry dated June 2, 2026, removes gpt-image-1-mini, gpt-image-1.5 and chatgpt-image-latest on Dec. 1, with the same two replacements. GPT-4o and GPT-4.1 snapshots, covered in our Oct. 13 article, follow on Jan. 14, 2027.
What the Replacements Cost
OpenAI’s model pages, read Oct. 23, price GPT-5.6 Sol at $4 per 1M input tokens and $20 per 1M output tokens. Terra costs $2 and $12. Sol and Terra are reasoning models with up to 128,000 output tokens and a Feb. 16, 2026 knowledge cutoff.
Our View
The substitute for gpt-3.5-turbo is Terra, a reasoning model.
We think developers calling gpt-3.5-turbo or o1 have two changes to make. Change the model name. Then check the new model’s price and settings against what your code assumes. If you run o1-pro, set the pro mode OpenAI lists for Sol.
The swap works as a drop-in only if the new model’s price and settings fit what you built around the old one.