Introduction to grok-build-0.1 and Grok Build
On this page
A model designed for code
grok-build-0.1 is xAI’s model optimized specifically for development in code editors. Unlike generalist models that try to do everything, this one focuses on a single mission: helping you write, understand, and transform code as quickly as possible. It is the fastest model in the lineup, in public beta on the API since June 1, 2026.
It replaces grok-code-fast-1, the previous code offering launched in August 2025: if you still see this name in configurations or tutorials, replace it with grok-build-0.1. xAI’s code offering now includes two components: the grok-build-0.1 model on the API, and the agentic development tool Grok Build, open source since July 15, 2026.
Why a model dedicated to code?
Generalist models like grok-4.5 are powerful, but their size and reasoning depth lead to higher latency and cost per token. For daily pair programming, you need an assistant that responds instantly, not a model that thinks at length before suggesting a change.
grok-build-0.1 solves this with three coherent design choices. Speed first: it’s the fastest model in the lineup, and suggestions arrive in real time — the experience feels like coding with a colleague sitting next to you, not submitting a ticket and waiting for a response. Cost next: at $1.00 per million input tokens and $2.00 output (under 200k context; $2.00 / $4.00 beyond), you can use it intensively all day without watching your consumption — a prerequisite for a code assistant to become a reflex rather than a resource you cut back on. Specialization finally: trained specifically for development tasks — completion, refactoring, debugging, test generation — it doesn’t need the reasoning depth of a generalist to excel in this scope, and that’s precisely what enables its speed and price.
The Grok Build tool
Around the model, xAI built Grok Build, an agentic development environment launched in beta on May 27, 2026 and open sourced on July 15. Its default model is grok-4.5, but you can switch to grok-build-0.1 to prioritize speed. Features include: the /goal command (high-level objectives, June 22), the Agent Dashboard (June 15), the Plugin Marketplace (June 11), and Workflows (July 23).
Grok Build and grok-build-0.1 are also integrated into Cursor, available on all plans.
Access via the xAI API
grok-build-0.1 is accessible via the xAI API at:
https://api.x.ai/v1This API is compatible with the OpenAI format, meaning any tool or editor supporting the OpenAI API can use grok-build-0.1 by simply changing the base URL and API key. You’ll see in the following lessons how to configure each editor.
Comparison with grok-4.5
This isn’t about replacing grok-4.5, but choosing the right tool for each situation:
- grok-build-0.1: fast implementation, code search, routine refactoring, boilerplate generation
- grok-4.5: complex debugging, system architecture, multi-step reasoning on difficult problems
In practice, you’ll use grok-build-0.1 for 80% of your daily interactions, and grok-4.5 when you’re stuck on a problem requiring deep analysis. Lesson 13 details this division of roles.
Pricing as of August 5, 2026 — prices change regularly: before any budget calculation, check the current rate card on the official xAI models and pricing page.
Key takeaways
- grok-build-0.1 is xAI’s fastest code model, in public API beta since June 1, 2026
- It replaces grok-code-fast-1, the previous code offering (2025)
- Pricing: $1.00 / $2.00 per million tokens under 200k context, with a 256k window
- Grok Build, the open source agentic tool (July 15, 2026), offers it alongside grok-4.5 (default model) and integrates with Cursor on all plans
- The API is OpenAI compatible (
https://api.x.ai/v1) - It complements grok-4.5 rather than replacing it