What are Databricks Foundation Model APIs for Atlas?

Databricks Foundation Model APIs allow Atlas, the terminal-native AI coding agent, to access leading large language models like Claude, GPT-5, and Gemini directly from within your Databricks lakehouse. This integration ensures that all model interactions are subject to your existing Unity Catalog governance, audit trails, and network boundaries, providing a secure environment for AI-powered development in 2026.

How do Databricks Foundation Model APIs enhance Atlas's security?

Databricks Foundation Model APIs enhance Atlas's security by serving models from inside your lakehouse. This means that every model call made by Atlas is governed by your Unity Catalog, recorded in your audit trail, and contained within your network boundary. An Atlas agent reading production schemas will not have that data leave your secure governance perimeter, which is critical for regulated teams.

What context window sizes are available with Databricks Foundation Model APIs?

Databricks Foundation Model APIs offer substantial context windows for Atlas. GPT-5.5 and GPT-5.4 models provide 1,050,000 tokens, while Claude Opus 4.7 and Claude Sonnet 4.6 offer 1,000,000 tokens. These large context windows enable Atlas to handle extensive codebases and complex development tasks effectively.

What is the pricing structure for Databricks Foundation Model APIs in Atlas?

The pricing for Databricks Foundation Model APIs in Atlas includes per-Mtok costs for input and output tokens, which sit on top of your DBU consumption. For example, Claude Opus 4.7 is $5.00/$25.00 per Mtok, GPT-5.4 is $2.50/$15.00 per Mtok, and GPT-5 Nano is $0.05/$0.40 per Mtok. It is important to note that the per-Mtok number is not your all-in cost, as DBU consumption is an additional factor.

Can I use open-source models with Databricks Foundation Model APIs in Atlas?

Yes, Databricks Foundation Model APIs support open-source models for Atlas. For instance, GPT OSS 120B is available at a highly competitive price of $0.072/$0.28 per Mtok, which is half the cost of using it on Groq. This provides a cost-effective option for developers who prefer open-weights models for their Atlas workflows.

What are the prerequisites for using Databricks Foundation Model APIs with Atlas?

To use Databricks Foundation Model APIs with Atlas, you must have an existing Databricks workspace. This is a non-negotiable prerequisite, as the API base URL is templated with your Databricks host. You will also need to export your `DATABRICKS_HOST` and `DATABRICKS_TOKEN` environment variables for Atlas to connect successfully.

How do I switch models within Atlas when using Databricks Foundation Model APIs?

Atlas allows you to switch the active model and provider on the fly. When using Databricks Foundation Model APIs, you can configure different models in your `atlas.json` file, for example, by setting `"model": "databricks/databricks-claude-sonnet-4-6"` for your primary coding slot or `"small_model": "databricks/databricks-gpt-5-nano"` for a cheaper slot. This flexibility enables you to leverage the specific strengths and cost profiles of various models served by Databricks.

Atlas with Databricks Foundation Model APIs in 2026

Atlas integrates direct with Databricks Foundation Model APIs, offering developers in 2026 a secure and governed environment for AI-powered coding. This integration ensures that your model calls, including those to Claude Opus 4.7 or GPT-5.5, remain within your Unity Catalog governance, audit trail, and network boundary. Pricing varies, with options like GPT-5 Nano at $0.05/$0.40 per Mtok and Claude Opus 4.7 at $5.00/$25.00 per Mtok, plus DBU consumption.

Why choose Databricks Foundation Model APIs for Atlas?

For developers in 2026, Databricks Foundation Model APIs offer unparalleled governance and security for Atlas, ensuring that sensitive code and data never leave your trusted environment. With models like Claude Opus 4.7 and GPT-5.5 offering context windows of 1,000,000 and 1,050,000 tokens respectively, your Atlas agent operates within your existing Unity Catalog controls.

Databricks Foundation Model APIs are ideal for Atlas users operating in regulated environments or those with strict data governance requirements. The core advantage is that Databricks serves leading frontier models such as Claude, GPT-5, and Gemini directly from inside your lakehouse. This means that your established Unity Catalog governance, comprehensive audit trail, and defined network boundary apply to every model call made by Atlas, just as they would to a standard table read. An Atlas agent that needs to read production schemas can do so without ever leaving your secure governance boundary. Furthermore, the catalog spans all three major frontier labs,Anthropic, OpenAI, and Google,alongside GLM-5.2 and Kimi K2.7 Code, providing a wide array of choices without requiring new vendor contracts for each model. Atlas, the terminal-native AI coding agent, lets you switch the active model and provider on the fly, making it easy to leverage the specific strengths of each model within your secure Databricks environment.

Cost and Context Window Tradeoffs with Databricks Foundation Model APIs

Understanding the cost structure is crucial for Atlas users, as Databricks Foundation Model APIs involve token prices on top of DBU consumption. For instance, Claude Opus 4.7 is priced at $5.00/$25.00 per Mtok, while GPT OSS 120B offers a highly competitive rate of $0.072/$0.28 per Mtok, providing a cost-effective option for many coding tasks.

When evaluating Databricks Foundation Model APIs for Atlas, developers must consider both the token prices and the additional DBU consumption, which means the per-Mtok number is not your all-in cost, unlike some other platforms. However, Databricks offers a wide range of models with varying price points and context windows. For high-performance tasks, Claude Opus 4.7 provides a 1,000,000 token context window at $5.00/$25.00 per Mtok, and GPT-5.4 offers 1,050,000 tokens at $2.50/$15.00 per Mtok. For more cost-sensitive operations, GPT-5 Nano is available at $0.05/$0.40 per Mtok, and GPT OSS 120B is notably priced at $0.072/$0.28 per Mtok on Databricks, which is half the Groq price. Claude Sonnet 4.6 is another option at $3.00/$15.00 per Mtok with a 1,000,000 token context, and Gemini 3 Flash is $0.50/$3.00 per Mtok. These diverse options allow Atlas users to balance capability, context window size, and cost effectively, switching models on the fly as needed for different coding tasks.

When to choose a different model for Atlas

While Databricks Foundation Model APIs offer significant advantages, particularly for regulated teams, they require an existing Databricks workspace. If you are not already a Databricks customer in 2026, the initial setup and overhead might make alternative model providers a more straightforward choice for driving Atlas.

Databricks Foundation Model APIs are specifically designed for integration within the Databricks ecosystem. A primary tradeoff is the requirement for a Databricks workspace. The base URL for the API is templated with your host (e.g., `https://${DATABRICKS_HOST}/ai-gateway/mlflow/v1`), which represents dead weight if your organization is not already a Databricks customer. For developers who do not have an existing Databricks infrastructure, the effort involved in setting up a workspace solely for Atlas model access might outweigh the benefits of the enhanced governance and audit capabilities. In such scenarios, other model providers that offer direct API access without a platform-specific prerequisite might provide a quicker and simpler path to integrating large language models with Atlas. Atlas lets you switch the active model and provider on the fly, so you can easily experiment with different options to find the best fit for your specific operational context and existing infrastructure.

Setup

01Export your Databricks workspace hostname: `export DATABRICKS_HOST="your-workspace-hostname"`
02Export your Databricks personal access token: `export DATABRICKS_TOKEN="your-personal-access-token"`
03Run `atlas models databricks` to confirm both environment variables resolved and the model catalog loaded successfully.
04Pin a high-context model for coding: Set `"model": "databricks/databricks-claude-sonnet-4-6"` in your `atlas.json` for the 1,000,000 token coding slot.
05Configure a cost-effective small model: Set `"small_model": "databricks/databricks-gpt-5-nano"` at $0.05/$0.40 per Mtok, or `"databricks/databricks-gpt-oss-120b"` at $0.072/$0.28 per Mtok for an open-weights cheap slot.

Frequently asked questions

What are Databricks Foundation Model APIs for Atlas?: Databricks Foundation Model APIs allow Atlas, the terminal-native AI coding agent, to access leading large language models like Claude, GPT-5, and Gemini directly from within your Databricks lakehouse. This integration ensures that all model interactions are subject to your existing Unity Catalog governance, audit trails, and network boundaries, providing a secure environment for AI-powered development in 2026.
How do Databricks Foundation Model APIs enhance Atlas's security?: Databricks Foundation Model APIs enhance Atlas's security by serving models from inside your lakehouse. This means that every model call made by Atlas is governed by your Unity Catalog, recorded in your audit trail, and contained within your network boundary. An Atlas agent reading production schemas will not have that data leave your secure governance perimeter, which is critical for regulated teams.
What context window sizes are available with Databricks Foundation Model APIs?: Databricks Foundation Model APIs offer substantial context windows for Atlas. GPT-5.5 and GPT-5.4 models provide 1,050,000 tokens, while Claude Opus 4.7 and Claude Sonnet 4.6 offer 1,000,000 tokens. These large context windows enable Atlas to handle extensive codebases and complex development tasks effectively.
What is the pricing structure for Databricks Foundation Model APIs in Atlas?: The pricing for Databricks Foundation Model APIs in Atlas includes per-Mtok costs for input and output tokens, which sit on top of your DBU consumption. For example, Claude Opus 4.7 is $5.00/$25.00 per Mtok, GPT-5.4 is $2.50/$15.00 per Mtok, and GPT-5 Nano is $0.05/$0.40 per Mtok. It is important to note that the per-Mtok number is not your all-in cost, as DBU consumption is an additional factor.
Can I use open-source models with Databricks Foundation Model APIs in Atlas?: Yes, Databricks Foundation Model APIs support open-source models for Atlas. For instance, GPT OSS 120B is available at a highly competitive price of $0.072/$0.28 per Mtok, which is half the cost of using it on Groq. This provides a cost-effective option for developers who prefer open-weights models for their Atlas workflows.
What are the prerequisites for using Databricks Foundation Model APIs with Atlas?: To use Databricks Foundation Model APIs with Atlas, you must have an existing Databricks workspace. This is a non-negotiable prerequisite, as the API base URL is templated with your Databricks host. You will also need to export your `DATABRICKS_HOST` and `DATABRICKS_TOKEN` environment variables for Atlas to connect successfully.
How do I switch models within Atlas when using Databricks Foundation Model APIs?: Atlas allows you to switch the active model and provider on the fly. When using Databricks Foundation Model APIs, you can configure different models in your `atlas.json` file, for example, by setting `"model": "databricks/databricks-claude-sonnet-4-6"` for your primary coding slot or `"small_model": "databricks/databricks-gpt-5-nano"` for a cheaper slot. This flexibility enables you to leverage the specific strengths and cost profiles of various models served by Databricks.

Try SeaShell in your terminal

The terminal-native AI coding agent. Free core, single binary.

Install SeaShell

Related guides

Audit a Repo with Parallel Subagents in Atlas (2026 Workflow)

How to audit a repo with parallel subagents in Atlas in 2026: the task tool launches explore subagents in their own sessions, so only conclusions return to your context.

Atlas vs Gemini CLI: Terminal AI Coding Agents in 2026

Comparing Atlas and Gemini CLI in 2026: Atlas offers terminal-native TUI, detailed change review, and BYO model flexibility. Gemini CLI provides a million-plus token context and a generous free tier.

Atlas vs Sourcegraph Cody: Terminal AI Coding Agents in 2026

Comparing Atlas and Sourcegraph Cody in 2026. Atlas offers a terminal-native TUI with permission-gated tool calls and local embeddings. Sourcegraph Cody excels in large enterprise monorepos with cross-repo search via

Atlas vs Mistral Vibe for Code: Terminal AI Coding Agents in 2026

Compare Atlas and Mistral Vibe for Code in 2026. Atlas offers terminal-native TUI, permission-gated tools, and local embeddings. Mistral Vibe provides a four-model stack and EU data sovereignty.

Atlas vs Magic.dev: Terminal AI Coding Agents in 2026

Compare Atlas, the terminal-native AI coding agent, with Magic.dev's research claims in 2026. Evaluate product availability, code safety, and pricing for developers.

Atlas for Java in 2026

Adopt Atlas, the terminal-native AI coding agent, for Java development in 2026. Enhance your workflow with intelligent code search, refactoring, and robust safety features for Maven and Gradle projects.

Atlas for Axum in 2026

Atlas is a terminal-native AI coding agent for Axum in 2026. It decodes tower trait-bound errors, adds IntoResponse types, and runs cargo nextest run.

Atlas vs Amp: Choosing Your Terminal AI Coding Agent in 2026

Compare Atlas and Amp, two leading terminal AI coding agents for 2026. Atlas offers a free core and TUI, while Amp features Oracle and remote Orbs with pay-as-you-go pricing.

Browse this resource hub