SAN FRANCISCO — xAI has made its Grok Build 0.1 coding model available to all developers through the xAI API in public beta, removing the subscription barrier that previously limited access to the tool and opening the agentic coding platform to the broader developer ecosystem.
A Coding Model Built for Agents
Grok Build 0.1 is a coding model trained specifically for agentic tasks — multi-step workflows where an AI needs to plan, search, and execute rather than respond to a single prompt. The model powers the Grok Build CLI that xAI shipped in May 2026, and developers can now access the same underlying model programmatically without a SuperGrok or X Premium+ subscription.
The model is served at over 100 tokens per second and supports a 256,000-token context window — large enough to hold an entire mid-size codebase in working memory. Pricing is set at $1 per million tokens for input and $2 per million for output, placing it competitively alongside similar offerings from Anthropic and OpenAI.
Eight Agents, One Codebase
One of Grok Build 0.1's distinguishing features is support for parallel agent execution. Developers can deploy up to eight agents simultaneously, each following a plan → search → build workflow. That architecture is designed for large-scale refactoring and feature work where sequential execution would be a bottleneck.
The model also supports Model Context Protocol (MCP), making it interoperable with the growing ecosystem of tools and servers built around that standard. For teams already using other agent-first tools, Grok Build 0.1 can integrate alongside existing workflows via the standard xAI API.
