OpenAI segments enterprise

OpenAI has split ChatGPT more sharply by customer type: some older model choices disappeared from the chat product, while enterprise plans gained more ways to ground answers in company data. (help.openai.com) OpenAI’s help center says that, as of February 13, 2026, GPT-4o, GPT-4.1, GPT-4.1 mini, o4-mini, and GPT-5 Instant and Thinking were retired from ChatGPT. The same notice says application programming interface, or API, access did not change. (help.openai.com) For ChatGPT Business, Enterprise, and Education customers, OpenAI also kept GPT-4o inside custom GPTs until April 3, 2026, even after removing it from the main ChatGPT model picker. Business and Enterprise model pages now list GPT-5.3 Instant as the default high-volume option. (help.openai.com, help.openai.com) A second change landed on the enterprise side: OpenAI says Business, Enterprise, and Education workspaces can now bring “company knowledge” into ChatGPT for answers tied to an organization’s own files and systems. The feature is presented as a way to keep users inside one chat instead of switching across internal tools. (help.openai.com) OpenAI’s documentation says these connected sources were previously called connectors and were renamed “apps” on December 17, 2025. In ChatGPT, those apps can search connected services and pull relevant context into a conversation. (help.openai.com, help.openai.com) The trust boundary is different by plan. OpenAI says data accessed from apps is not used to train models by default for Business, Enterprise, and Education customers, while consumer plans can allow training from app data if the “Improve the model for everyone” setting is on. (help.openai.com) The company has also added more admin controls around those workspaces. Recent Business release notes say workspace owners and admins can manage app actions through a simplified control model, while Enterprise and Education release notes say apps with sync now support in-region data residency in Japan. (help.openai.com, help.openai.com) That leaves a clearer split in OpenAI’s product stack. ChatGPT is narrowing the set of in-product model choices, while paid workspaces are getting more controls over which models employees can use and which internal sources can shape an answer. (help.openai.com, help.openai.com) The practical result for large organizations is less about one model name disappearing than about where decisions now sit: in admin settings, custom GPTs, data residency rules, and connected knowledge sources. OpenAI’s recent updates show those levers moving further into Business, Enterprise, and Education plans. (help.openai.com, help.openai.com, help.openai.com)