OpenAI tweaks model lineup

OpenAI quietly changed the names ordinary ChatGPT users see, and a bunch of familiar options disappeared on February 13, 2026: GPT-4o, GPT-4.1, GPT-4.1 mini, OpenAI o4-mini, and GPT-5 Instant and Thinking were all retired from the ChatGPT picker. (help.openai.com) What replaced them is simpler on the surface and messier underneath: ChatGPT now centers on GPT-5.3 and GPT-5.4, while OpenAI says API access remains unchanged for developers and enterprise customers. (help.openai.com) That split matters because ChatGPT and the application programming interface are now being treated like two different storefronts for the same engine room: one is a consumer menu with fewer labels, and the other is a contract product where stability matters more than branding. (help.openai.com, developers.openai.com) OpenAI’s own help pages make the divide explicit: retired ChatGPT models are gone from the app, but Business, Enterprise, and Edu customers kept GPT-4o inside custom GPTs until April 3, 2026. (help.openai.com) The migration also reaches user-built bots. OpenAI says GPTs that were pinned to retired models are automatically moved to the closest GPT-5.3 Instant or GPT-5.4 Thinking and Pro equivalent instead of breaking outright. (help.openai.com) Under the hood, OpenAI is pushing GPT-5.4 as the new default workhorse in the application programming interface, and its developer docs say to start there for general work and most coding tasks. (developers.openai.com, developers.openai.com) OpenAI’s March 5, 2026 launch post tied GPT-5.4 to professional use, not casual chat, and said the model ships across ChatGPT, the application programming interface, and Codex at the same time. (openai.com) Codex is OpenAI’s coding agent inside ChatGPT, and when it launched in May 2025, the company said it could write features, fix bugs, answer questions about a codebase, and run tests over sessions lasting from one minute to 30 minutes. (techcrunch.com) Now the pricing ladder appears to be moving too. TechCrunch and The Economic Times both reported on April 9 and April 10, 2026 that OpenAI is testing a $100 Pro plan for Codex with five times the Codex usage of the $20 Plus tier. (techcrunch.com, economictimes.indiatimes.com) Put together, the pattern is pretty clear: fewer model names in ChatGPT, more continuity in the application programming interface, and a separate push to charge heavy coding users for extra capacity instead of making every customer learn OpenAI’s internal model family tree. (help.openai.com, developers.openai.com, techcrunch.com)