OpenAI fashions, together with GPT-5.1, are actually typically accessible on the Databricks Information Intelligence Platform. This launch is a key a part of our Week of Brokers, showcasing new releases that can assist you construct, govern, and deploy highly effective AI brokers securely at scale.
With GA, OpenAI fashions on Databricks attain full manufacturing readiness. This marks one other step towards Databricks’ mission to make main basis fashions accessible natively inside a single, ruled platform so enterprises can construct high-quality brokers that cause over their information securely and at scale.
Construct with OpenAI, The place Your Information Lives
Now, enterprises can convey OpenAI’s frontier intelligence on to their ruled information, with out shifting information round. This integration permits methods that not solely generate language but in addition perceive, cause, and act on enterprise information in actual time.
Groups are already utilizing this mix to:
- Enrich hundreds of thousands of help tickets or product critiques with automated insights
- Generate structured summaries and motion objects from assembly transcripts or contracts
- Construct conversational brokers that reply questions grounded in reside Delta information
What You Can Do with OpenAI Fashions on Databricks?
Discover Fashions within the AI Playground
The AI Playground gives the quickest path to begin experimenting. Open the Playground tab, select an OpenAI mannequin, and examine GPT-5.1, GPT-5, GPT-mini, and GPT-nano aspect by aspect. Modify temperature, max tokens, and different parameters interactively, then export the very best immediate immediately right into a pocket book or SQL cell.
Apply OpenAI Fashions to Your Information
OpenAI fashions can be utilized like a built-in operator in SQL or Python, making it easy to investigate contracts, PDFs, transcripts, or pictures immediately the place your information lives. Once you run these queries, Databricks mechanically scales OpenAI mannequin capability within the backend to deal with every thing from a handful of rows to hundreds of thousands, making certain quick, dependable outcomes with out additional setup.
Automate Doc Intelligence
OpenAI fashions might be chained into Lakeflow Declarative Pipelines to construct GenAI ETL, ingesting information from methods like Dynamics 365, making use of AI transformations reminiscent of summarization or classification, and orchestrating the stream with built-in governance and observability. This sample scales to hundreds of thousands of paperwork, creating end-to-end document-intelligence pipelines with lineage, auditability, and high quality monitoring in-built.
Construct Clever Brokers with OpenAI and Agent Bricks
Agent Bricks turns OpenAI fashions into clever brokers by connecting them to enterprise information within the Lakehouse, vectorized data via Vector Search, and exterior methods by way of MCP.
As soon as linked, Agent Bricks automates immediate and mannequin optimization, selecting the right mixture for every use case and repeatedly enhancing efficiency via built-in analysis, observability, and governance.
When prepared, brokers might be securely deployed at scale, totally ruled by Unity Catalog and monitored via AI Gateway.
Govern and Monitor with AI Gateway
All OpenAI site visitors on Databricks is managed via AI Gateway, the unified management aircraft for enterprise AI workloads. Each GPT-5.1 request is ruled, logged, and auditable — totally built-in with Unity Catalog. Groups acquire visibility into utilization and token spend, can implement charge limits, and allow fallbacks to make API calls extra dependable.
