Snowflake lets its Cortex AI gateway choose models itself

Snowflake is updating its Cortex AI gateway so it can choose models on its own, according to Techzine Global, a move aimed at lowering the barrier to building enterprise AI applications.

Traditionally, development teams have had to manually pick and configure the right model for each task. With this update, the Cortex AI gateway can handle model selection automatically, leaving routing decisions to the platform.

Cortex is the AI layer inside Snowflake's data cloud, and the AI gateway sits between applications and the underlying models, managing calls, routing, and governance. Letting the gateway pick models means enterprises can focus more on business logic.

The value of automatic model selection is workflow simplification: companies no longer need to track every model version and capability change, since the gateway can route tasks to suitable models while keeping a unified governance entry point.

The open question is how automatic selection interacts with cost control and evaluation — if the gateway can make decisions based on task type and budget, it could substantially reduce the management overhead of AI applications.

What to watch next is whether Snowflake reveals more about the selection criteria, the range of supported models, and enterprise governance details, which will determine how quickly the feature is adopted in production.