Which Grok models are available on Databricks Agent Bricks?

Grok on Databricks ships with two xAI models: Grok 4.3, the reasoning flagship with a one-million-token context window and a December 2025 knowledge cutoff, and Grok Build 0.1, the coding-specialised variant. Both run inside Agent Bricks alongside OpenAI, Anthropic, Gemini, Qwen, and Kimi options.

How does Grok on Databricks handle enterprise data?

Grok on Databricks connects directly to context stored in the Databricks Lakehouse. Agents reason over a company's structured and unstructured data inside Databricks' governance layer, so prompts and retrieval results stay in the customer's account rather than being routed through external SaaS pipelines.

How much does Grok 4.3 cost when called through Databricks?

On xAI's published API price list Grok 4.3 is $1.25 per million input tokens and $2.50 per million output tokens, and Grok Build 0.1 is $1.00 in / $2.00 out per million tokens. Databricks-tier pricing follows that schedule, with billing details surfaced in the Agent Bricks marketplace listing.

Where else can teams run Grok besides Databricks?

xAI's announcement points to Grok on Amazon Bedrock as the other governed cloud-platform path, alongside the direct xAI API. Databricks is the second major data-plus-AI platform to host Grok, after Bedrock added it earlier in the week, broadening enterprise access for SpaceXAI's frontier models.

xAI · 2026-06-18 · major

Grok on Databricks — xAI models land in Agent Bricks via SpaceX deal

Grok on Databricks makes Grok 4.3 and Grok Build 0.1 natively callable from Databricks Agent Bricks, so enterprise teams can wire xAI models into governed Lakehouse data without external pipelines.

xAI announcement card for Grok models on Databricks Agent Bricks

xAI's Grok 4.3 and Grok Build 0.1 are now native model options inside Databricks' Agent Bricks platform.

Key specs

Context window	1M tokens
Grok 4.3 price	$1.25 / $2.50 per 1M tokens

Quick facts

Announced	Databricks Data + AI Summit 2026
Models	Grok 4.3, Grok Build 0.1
Platform	Databricks Agent Bricks
Data path	Native Lakehouse context, no external pipeline
Partner entity	SpaceXAI
Sibling availability	Grok also on Amazon Bedrock

What is it?

Grok on Databricks adds two xAI models — Grok 4.3 reasoning and Grok Build 0.1 coding — as native options inside Databricks Agent Bricks, Databricks' developer platform for production agents. The integration was announced at the Databricks 2026 Data + AI Summit through a partnership with SpaceXAI. Agent Bricks teams can now pick Grok from the same menu as OpenAI, Anthropic, Gemini, Qwen, and Kimi.

How does it work?

Inside Agent Bricks, Grok models read context directly from Databricks' Lakehouse rather than from an external retrieval service, so structured tables and unstructured documents stay in the customer's governed environment. xAI says agents can reason over large volumes of enterprise data without routing it through outside pipelines, and notes that model partners do not retain data submitted through these features.

Why does it matter?

For Databricks customers this is the first time xAI's frontier models are a one-click choice next to OpenAI, Anthropic, and Google in a governed enterprise platform. It matters because most regulated organisations cannot move data into a vendor's hosted endpoint, and Grok on Databricks lets them try Grok 4.3's one-million-token context against their own Lakehouse tables without that hurdle.

Who is it for?

enterprise engineering teams already building on Databricks Agent Bricks

Frequently asked questions

Which Grok models are available on Databricks Agent Bricks?: Grok on Databricks ships with two xAI models: Grok 4.3, the reasoning flagship with a one-million-token context window and a December 2025 knowledge cutoff, and Grok Build 0.1, the coding-specialised variant. Both run inside Agent Bricks alongside OpenAI, Anthropic, Gemini, Qwen, and Kimi options.
How does Grok on Databricks handle enterprise data?: Grok on Databricks connects directly to context stored in the Databricks Lakehouse. Agents reason over a company's structured and unstructured data inside Databricks' governance layer, so prompts and retrieval results stay in the customer's account rather than being routed through external SaaS pipelines.
How much does Grok 4.3 cost when called through Databricks?: On xAI's published API price list Grok 4.3 is $1.25 per million input tokens and $2.50 per million output tokens, and Grok Build 0.1 is $1.00 in / $2.00 out per million tokens. Databricks-tier pricing follows that schedule, with billing details surfaced in the Agent Bricks marketplace listing.
Where else can teams run Grok besides Databricks?: xAI's announcement points to Grok on Amazon Bedrock as the other governed cloud-platform path, alongside the direct xAI API. Databricks is the second major data-plus-AI platform to host Grok, after Bedrock added it earlier in the week, broadening enterprise access for SpaceXAI's frontier models.

Try it

https://www.databricks.com/blog/agent-bricks-dais-2026