Model API
A web service that lets programs send prompts to a hosted model and get answers back, usually paid per token.
Also: API, inference API
Most people and companies use powerful models through an API rather than running them. The developer sends a request with the prompt and settings and receives the generated text, priced per million input and output tokens. OpenAI opened its API with GPT-3 in 2020; today Anthropic, Google, Mistral, DeepSeek, Alibaba and others offer their own, and cloud platforms and inference companies host many models behind one interface. Many providers copy OpenAI's request format so that switching models takes little code.