OpenAI trims ChatGPT models
OpenAI has removed older GPT variants from ChatGPT’s consumer interface while keeping API access intact and preserving GPT‑4o for business customers inside custom GPTs, signaling clearer consumer/enterprise product segmentation. The company also expanded Codex with a $100 Pro plan aimed at heavy coding users, a move that tightens monetization and tiers for professional developers. (help.openai.com) (economictimes.indiatimes.com)
OpenAI quietly did two different cleanups at once: it removed a stack of older model choices from the ChatGPT app on February 13, 2026, and it left those same models available through the application programming interface, which is the developer pipe companies use to plug models into their own software. The models cut from ChatGPT included GPT-4o, GPT-4.1, GPT-4.1 mini, OpenAI o4-mini, and GPT-5 Instant and GPT-5 Thinking. OpenAI’s help center says the consumer chat product now centers on GPT-5.3 and GPT-5.4 instead. That split matters because ChatGPT and the application programming interface are now being treated like two different stores. One is a simplified shelf for everyday users, and the other is a parts catalog for developers who still want older behavior or compatibility. OpenAI briefly kept one exception alive for paying organizations. ChatGPT Business, Enterprise, and Edu customers could still use GPT-4o inside custom GPTs until April 3, 2026, even after regular ChatGPT users lost it. A custom GPT is OpenAI’s built-in way to make a specialized chatbot with saved instructions, tools, and files. Keeping GPT-4o there a little longer gave companies time to swap out internal assistants without breaking workflows overnight. OpenAI is also redrawing the price ladder around coding. This week it added a new $100-a-month ChatGPT Pro tier aimed at people who use Codex heavily, placing it between the $20 Plus plan and the $200 Pro plan already on the market in earlier pricing coverage. Codex is OpenAI’s coding agent inside ChatGPT, which means it can read code, reason through changes, and take actions across software projects instead of just answering a one-off question. OpenAI’s current pricing pages list Codex access across paid plans, but the new $100 offer raises the usage ceiling for people who spend hours in it. Reports on the launch say the $100 plan gives 5 times the Codex usage of Plus, with a temporary 10 times boost running through May 31. That is a classic software move: keep the casual plan cheap, charge the heavy users more, and leave the very top tier for the people who want the biggest limits. Put those two changes together and the shape of the business gets clearer. ChatGPT is being trimmed into a cleaner product with fewer model names, while developers and companies are being pushed toward separate lanes for legacy access, custom tools, and higher-usage coding work. That is why this does not look like a simple model retirement. It looks like OpenAI deciding that consumers should see fewer knobs, businesses should get managed transition windows, and professional coders should pay for the extra electricity and compute time they burn through.
