OpenAI has reportedly cut the cost of serving guest ChatGPT users by more than half. The reduction points to continuing efficiency gains in inference, routing and serving infrastructure.
Lower response costs matter because consumer AI products operate at enormous scale. Even small savings per interaction can change the economics of free access, onboarding and experimentation.
For the broader market, the report shows why model capability is only part of the competition. The companies that can serve useful answers cheaply have more room to grow usage without immediately raising prices.