OpenAI launches ChatGPT Enterprise

OpenAI launched ChatGPT Enterprise, a business tier sitting above the existing $20-a-month Plus subscription. It offered unlimited, higher-speed access to GPT-4, a context window extended to 32,000 tokens (roughly 25,000 words, up from Plus’s smaller default), unlimited use of the Advanced Data Analysis tool (formerly Code Interpreter), an admin console with single sign-on and usage dashboards, and shareable internal conversation templates.

The pitch to businesses centred on data handling rather than capability: OpenAI said it would not train its models on Enterprise customers’ conversations or business data, and that conversations would be encrypted both in transit and at rest, with SOC 2 compliance. This addressed the most commonly cited reason large organisations had held back from ChatGPT adoption — the risk of proprietary or client data leaking into a model OpenAI might later train on. Pricing was not published; OpenAI said cost would depend on a company’s usage and use case, to be negotiated directly.

OpenAI said the product had been built over roughly a year with input from more than 20 companies of varying sizes, and named early users including Block, Canva, Carlyle, Estée Lauder, PwC and Zapier. The company also claimed ChatGPT was already in use, in some form, at over 80% of Fortune 500 companies — Enterprise was framed as formalising and securing usage that was largely already happening informally.

The launch mattered less for any new capability than for what it signalled about OpenAI’s business model: a consumer chatbot phenomenon converting into a recurring enterprise software line, competing directly with the seat-based licensing Microsoft was building around GPT-4 through Copilot. It also set a template — unlimited access, security guarantees, admin controls — that Anthropic, Google and other vendors would each replicate with their own enterprise tiers within the following year.