We are releasing new models, reducing prices for GPT‑3.5 Turbo, and introducing new ways for developers to manage API keys and understand API usage. The new models include: