OpenAI, Databricks pair GPT models with governed data via Genie Ontology

The partnership builds on a multi-year agreement that began in September 2025, with each company acting as a customer of the other. OpenAI is one of Databricks' largest customers, and every Databricks engineer is getting access to Codex, according to the post. Databricks gets early access to new OpenAI models, and OpenAI models lead on OfficeQA, a Databricks benchmark for real document work.

Three starting patterns are outlined: accelerating work for business users with ChatGPT and the Genie plugin, coding with enterprise-level governance using OpenAI Codex routed through Databricks Unity Gateway, and custom agents built with OpenAI models on Databricks Model Serving and deployed as Databricks Apps. The Genie plugin runs on the Genie MCP, and Genie Ontology stores approved metrics and business definitions so ChatGPT applies the same definitions as finance reports. For Codex, a Databricks credential signs developers in while the OpenAI key stays in the Gateway, and every request is logged, letting engineering leaders set budgets and review spend by team.

Additional integrations include the Neon plugin for Codex, Databricks Agent Skills for Codex, a Codex Data Science plugin, a ChatGPT Data Analyst plugin, and a one-click ChatGPT connection that applies each user's existing permissions without admin setup. Teams already using ChatGPT and Codex can start with the Genie plugin and Unity Gateway, while teams building in Databricks can start with OpenAI GPT models there.