OpenAI trims ChatGPT lineup

OpenAI has retired a number of older ChatGPT models from the ChatGPT consumer interface while keeping API access stable for enterprises, a move that separates a simpler front‑end from enterprise continuity. The company’s help centre documents the removals, and there are market rumours that OpenAI’s next internal model, codenamed “Spud,” finished pretraining in late March and may ship soon (help.openai.com, abhs.in).

OpenAI has removed several older models from the ChatGPT app and website, while leaving application programming interface access in place for developers and business customers. (openai.com) The company said GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini were retired from ChatGPT on February 13, 2026, alongside the previously announced removal of GPT-5 Instant and GPT-5 Thinking from the consumer interface. OpenAI’s help center says those models remain available in the application programming interface. (help.openai.com, help.openai.com) OpenAI also gave ChatGPT Business, Enterprise, and Edu customers a temporary bridge: GPT-4o stayed available inside Custom GPTs until April 3, 2026. After that date, the help center says those retired models were no longer available in ChatGPT at all. (help.openai.com, help.openai.com) The change leaves ChatGPT with a narrower front-end lineup built around newer GPT-5 variants and automatic model switching. OpenAI’s model notes say GPT-5 is presented as a single system in ChatGPT, while paid users still get a model picker for GPT-5 and GPT-5 Thinking. (help.openai.com) OpenAI has framed the removals as a simplification drive rather than a platform rollback. In a February 13 post, the company said retiring older options lets it focus on “the models most people use today,” and it noted that the application programming interface was unchanged. (openai.com) That split matters for customers who use ChatGPT as a consumer product versus companies that have built software on top of OpenAI’s back-end. The chat interface lost legacy choices in February, but enterprise continuity was preserved through the application programming interface and, for a few extra weeks, through Custom GPTs. (help.openai.com, help.openai.com) The cleanup also fits a product plan OpenAI signaled earlier. In February 2025, Chief Executive Sam Altman said the company wanted to simplify its offerings and move away from a long list of separate model names. (techmeme.com) There is also fresh speculation about what comes next. A widely circulated report summarized by several secondary outlets says OpenAI’s internal next-model codename is “Spud” and that pretraining finished on March 24, 2026, but OpenAI has not confirmed those details in its help center or product posts. (abhs.in, humai.blog) For now, the concrete change is simpler than the rumor cycle: as of April 12, 2026, older models are gone from ChatGPT, and OpenAI’s official documentation says the application programming interface remains unchanged. (help.openai.com, openai.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.