Databricks Free Edition Plans

The plan supports experimentation with AI models and foundation models, AI applications, and agents, but Databricks does not publish a single exhaustive list of exact model identifiers for Free Edition on the cited plan pages. Some models and serving configurations are unavailable or limited.

Serverless-only environment with limited compute size and usage. One SQL warehouse limited to 2X-Small, maximum five concurrent job tasks per account, one active Lakeflow pipeline per pipeline type, one AI Search endpoint with one search unit, and up to three Databricks Apps. Exceeding quotas can shut down compute until the quota resets.

For personal, learning, training, and other non-commercial use only. One serverless workspace and one metastore per account. No classic compute, custom compute configurations, production support, SLA, account console, account-level APIs, SSO, SCIM, compliance enforcement, security customization, private networking, custom workspace storage locations, R or Scala, clean rooms, online tables, or legacy Databricks features. No GPU serving endpoints, provisioned throughput, custom GPU or batch-inference models, or Marketplace provider capability. Databricks may delete accounts inactive for a prolonged period.

Provided at no cost with no payment required. Usage is subject to fair-use quotas and resource limits.

The Enterprise tier provides access to the Databricks platform's model-serving, AI application, and governance capabilities, but it does not represent an unlimited bundle of named models. Exact model availability depends on the selected Databricks service, cloud, region, deployment configuration, and any external model-provider connection. Model inference and serving usage may be billed separately.

Usage is not provided as a fixed user quota. Capacity and cost depend on the Databricks products, compute, storage, serverless services, model-serving workloads, and other platform usage selected by the customer. Enterprise resource limits are documented separately and some limits may be increased through the account team.

Enterprise is a commercial platform tier rather than an individual flat-rate AI assistant subscription. It does not provide a universal fixed allowance for users, messages, files, or model calls. Availability and pricing can vary by cloud provider and by Databricks service. Some capabilities may be previews, region-dependent, product-specific, or require additional configuration or separate usage charges.

No fixed public subscription price is advertised. Databricks charges for platform products and compute usage, with committed-use contracts and enterprise pricing available through sales. Pricing varies by cloud provider, workload, products, and usage.

Databricks does not publish a single fixed model bundle for the Premium tier. Exact model and AI feature availability depends on the workspace configuration, cloud, product service, release status, and applicable usage charges.

Usage is not expressed as a fixed per-user allowance. Charges are based primarily on Databricks Units and other usage meters. Workspace and account quotas apply; for example, Databricks documents default workspace limits that differ between Premium and Enterprise tiers.

Advanced Enterprise-only security capabilities include customer-managed keys for certain encryption functions and encryption of traffic between cluster worker nodes, queries, query history, and query results. Pricing and availability of individual workloads can vary by cloud provider and region. Cloud infrastructure, storage, networking, and related charges may be billed separately by the cloud provider.

No fixed public subscription price. Databricks Premium is billed primarily according to platform usage, including compute and related services, with pricing varying by cloud provider, region, workload, and commercial agreement. Pay-as-you-go and committed-use arrangements are available.