OpenAI clarifies enterprise stack
- OpenAI refined ChatGPT product segmentation, keeping ads off paid tiers and adding a 'company knowledge' feature for Business, Enterprise and Education plans. - It removed some ChatGPT models, retained GPT-4o for Business/Enterprise custom GPTs, and published explicit Codex pricing and credits. - The updates formalise plan-based model access and usage economics even as rumours of a GPT-5.5 leak circulate without official confirmation. (help.openai.com; help.openai.com; moneycontrol.com)
OpenAI has redrawn the lines between its workplace ChatGPT plans, adding “company knowledge” to Business, Enterprise, and Education while keeping ads off paid tiers. (help.openai.com; help.openai.com) The new company knowledge feature lets ChatGPT search approved workplace apps and answer with citations back to the original source; it is available on ChatGPT Web, not the desktop or mobile apps. For Enterprise and Education, admins must enable apps first; for Business, apps are enabled by default. (help.openai.com) OpenAI also formalized a two-seat structure for workplace accounts on April 2: standard ChatGPT seats, which include ChatGPT and Codex, and Codex-only seats, which are usage-based and do not include ChatGPT access. Business standard seats are listed at $25 per user monthly or $20 annually in most countries, with a two-seat minimum. (help.openai.com; help.openai.com) That split turns ChatGPT’s workplace plans into something closer to a software bundle with metered extras. OpenAI’s Business and Enterprise help pages now describe baseline access, workspace credits, usage visibility, and spend controls alongside security features such as single sign-on and System for Cross-domain Identity Management, or SCIM. (help.openai.com; help.openai.com) The pricing language has shifted too. OpenAI’s Codex rate card says that, as of April 2, 2026, new and existing Plus, Pro, and Business customers, plus new Enterprise customers, are billed on token usage rather than per-message estimates, with charges expressed as credits per 1 million input, cached input, and output tokens. (help.openai.com) OpenAI has also been pruning the model menu inside ChatGPT. Its help center says GPT-4o, GPT-4.1, GPT-4.1 mini, OpenAI o4-mini, and GPT-5 Instant and Thinking were retired from ChatGPT on February 13, 2026, with GPT-5.1 models retired on March 11; API access for those models remains unchanged. (help.openai.com) For workplace customers, GPT-4o got a short extension inside custom GPTs. OpenAI said Business, Enterprise, and Education users could keep GPT-4o in custom GPTs until April 3, 2026, after which it would be fully retired across all plans. (help.openai.com; help.openai.com) The ad policy draws another clean boundary. OpenAI’s release notes say ads are rolling out only to Free and Go users in markets including the United States, Australia, New Zealand, and Canada, while Plus, Pro, Business, Enterprise, and Education plans remain ad-free. (help.openai.com; help.openai.com) Rumors of a “GPT-5.5” have circulated this week, driven by leaks and speculation in tech outlets, but OpenAI’s public help center and release notes do not list an official GPT-5.5 launch as of April 23, 2026. The documents OpenAI has published instead focus on plan boundaries, model retirements, and how enterprise customers are billed when employees use more AI. (moneycontrol.com; help.openai.com; help.openai.com)
