OpenAI has begun offering some large customers a new payment model in recent months, according to The Information. Under the arrangement, customers pay only after AI completes a task, rather than being billed by tokens, usage, or seats.
The model is already being used in customer service, where companies can pay based on the AI’s completed support tasks. OpenAI has not disclosed specific customers, pricing, or how it defines a completed task, and it declined to comment. The company’s public Enterprise pricing still relies mainly on credits and token-based billing, suggesting the outcome-based approach is being negotiated separately with select large clients.
In January, OpenAI CFO Sarah Friar said AI business models could eventually shift toward charging for results. In July, OpenAI also argued that the more relevant measure is how much it costs to get an AI system to finish a task, and that idea is now appearing in real contracts.