OpenAI outlines a full-stack approach to making AI more affordable
OpenAI AI pricing cuts make advanced models cheaper, faster, and more useful See how new efficiency gains and flexible options can boost adoption and value
OpenAI published a July 31, 2026 company post describing how it is trying to make advanced AI more capable, cheaper, and more widely useful. The article says the company sees lowercost intelligence as a way to broaden adoption while also supporting more investment in research and infrastructure.
The post highlights recent price cuts for GPT5.6 Luna and GPT5.6 Terra, along with faster processing options for GPT5.6 Sol. OpenAI says these changes are meant to give customers more flexibility in balancing speed, reliability, intelligence, and cost.
The company also says efficiency gains come from more than just larger data centers. It points to improvements in routing, context management, speculative decoding, and serving software that reduced costs and improved tokengeneration efficiency. OpenAI argues that its work across infrastructure, models, platform, and products helps each layer improve the others, and it says demand from ChatGPT, ChatGPT Work, Codex, and the API informs its capacity planning.