databricks Platform · Updated Apr 21, 2026

CPU_MEDIUM and CPU_LARGE workload types for Model Serving (Beta)

The CPU MEDIUM and CPU LARGE workload types for Model Serving are now in Beta.

What it does

The CPU MEDIUM and CPU LARGE workload types for Model Serving are now in Beta. These workload types let you trade concurrency for more memory per worker on the same CPU hardware. Use them when your model needs more memory than the standard CPU workload type provides.

Why it matters

Doing so reduces risk of interruption for endpoints that are in use.

Key capabilities

defaults

python=3.8.8

mlflow

scikit-learn==0.23.2

Recommended use cases

For high QPSlow latency use casesroute optimization is the optimal

Architecture

Databricks Platform → Platform

In short

The CPU MEDIUM and CPU LARGE workload types for Model Serving are now in Beta. These workload types let you trade concurrency for more memory per worker on the same CPU hardware.

Read the full documentation ↗

Key takeaway

The CPU MEDIUM and CPU LARGE workload types for Model Serving are now in Beta.

2026-04-21-cpu-medium-and-cpu-large-workload-types-for-model-serving-beta Generated by the Databricks Updates Agent · browse all updates