OpenAI to retire legacy ChatGPT models as it consolidates on GPT‑5 family

OpenAI will remove several older models from the ChatGPT product on Friday, Feb. 13, 2026, a change the company says will let it concentrate development on the models most users employ. Multiple outlets reporting on an OpenAI support notice and company blog named GPT‑4o, GPT‑4.1, GPT‑4.1 mini and OpenAI o4‑mini among the models slated for removal; one report also listed GPT‑5 as affected, an inconsistency in public accounts of the exact list.

OpenAI framed the step as difficult but necessary. As reported by CNET from the company blog, OpenAI said, "We know that losing access to GPT‑4o will feel frustrating for some users, and we didn't make this decision lightly," and added that, "Retiring models is never easy, but it allows us to focus on improving the models most people use today." The company further told reporters that only 0.1 percent of its users regularly use GPT‑4o to run tasks; CNET calculated that share at roughly 800,000 users based on OpenAI's 2025 enterprise report figure of about 800 million weekly active users.

The announcement follows a history of churn around GPT‑4o. Launched in May 2024, GPT‑4o earned a reputation for an approachable, multimodal conversational style that blended text, images and voice. When OpenAI released GPT‑5 last year it briefly removed GPT‑4o from the model picker, prompting a "huge and immediate backlash" and the company subsequently restored access for paid users. CNBC and 9to5Mac noted that CEO Sam Altman pledged on X to give "plenty of notice" before any retirement.

Observers say the retirements are part of a larger product consolidation into the GPT‑5 family, with personality and configuration features being folded into newer releases. ITP.net reported that OpenAI is embedding personality controls directly into GPT‑5.2, and that the move reflects a shift away from maintaining separate models for different user preferences. CNET also highlighted the emergence of GPT‑5.1 and GPT‑5.2 as the models to watch as replacements for legacy behavior patterns.

The change is limited to ChatGPT's product interface, and outlets including ITP.net and CNBC said OpenAI confirmed "there are no changes to the API at this time," a reassurance aimed at developers planning migrations. Still, users and developers who favored the older models are likely to test whether the new GPT‑5 variants reproduce the tone and features they valued.

The move has prompted renewed discussion about how companies balance product streamlining with niche user needs. CNET captured a strand of that debate with a rueful line: "Pour out a cold one for ChatGPT‑4o." Other commentators have warned that the conversational warmth that made GPT‑4o popular can also tip into problematic "AI sycophancy" — "when models are overly affectionate, becoming digital yes‑men that potentially validate users' dangerous ideas," a concern reflected in recent reporting.

With Feb. 13 set as the deadline for the ChatGPT interface change, users loyal to retiring models will face a choice: migrate to the newer GPT‑5 series or seek alternatives. OpenAI says the shifts will let it focus on the experiences most people use today, even as a vocal subset of users and safety experts scrutinize what is lost in the consolidation.