OpenAI GPT-6 Astra is available on Foundation Model APIs

The OpenAI GPT-6 Astra model is now available in Databricks for enterprise workloads requiring advanced reasoning, structured document processing, and agentic searches. It accepts multimodal inputs such as text and images and can be accessed directly through Unity Gateway.### Use casesImagine you need to process complex contracts in batches and perform detailed analyses combining text and images. With GPT-6 Astra integrated into your environment, you gain native access to a frontier model without managing custom infrastructure. What you can do:- Process and extract structured documents- Handle reasoning and coding tasks- Run agent-driven searches- Process long-context workloads### How to useYou can send requests to the model using its Unity Gateway model service name or via pay-per-token endpoints in your workspace.pythonimport mlflowfrom databricks.sdk import WorkspaceClientw = WorkspaceClient()response = w.serving_endpoints.query( name="databricks-gpt-6-astra", logged_requests=, messages=[{"role": "user", "content": "Analyze this document and list the main clauses."}])print(response.choices[0].message.content)What happens in this code:1. The Databricks SDK client initializes the workspace connection.2. The serving endpoint query method sends the prompt to the databricks-gpt-6-astra endpoint.3. The model processes the request using its reasoning capabilities and returns the structured answer.### Limitations and requirements- The model is available in pay-per-token mode.- Customers are responsible for ensuring compliance with applicable OpenAI model terms.