OpenAI API

The OpenAI API is the developer platform for building applications and agents on OpenAI's models. It gives programmatic access to frontier models with multimodal capabilities, along with tools for building agent workflows, grounding them in context and connecting them to business systems. The platform lists its current models with their specifications. Astra has a context length of about 1.05 million tokens, a 128K output limit and a knowledge cutoff of April 2026. Sol has the same context and output limits and a matching cutoff, and Luna has the same limits with a cutoff of May 2026. Builders work with the Agents SDK and the Responses API, add voice with the Live model, ground agents through built in tools such as web search, file search and remote MCP servers, and let them act across business systems. The site highlights use cases in coding, customer support, personalized recommendations and research, and offers prompting guidance and migration guides. It is aimed at engineers and product teams, from individual developers to enterprises, who want to embed language, voice and agent capabilities in their own software, with a sales contact route for larger deployments. Pricing is per million tokens at standard processing rates for context lengths under 272K. Astra costs $10.00 for input and $50.00 for output, Sol costs $2.00 for input and $10.00 for output, and Luna costs $0.10 for input and $0.50 for output. Batch processing and regional data residency pricing are documented separately.