Grok Build 0.1

Grok Build 0.1 is xAI's specialized coding model built for agentic software development, capable of planning, writing, refactoring, and iterating on code across multi-step processes. xAI trained it specifically for coding work such as web development, debugging, and Model Context Protocol support, and it is the same model that powers the Grok Build product. The model handles a 256,000-token context window, giving it room to maintain large codebases and lengthy development sessions without losing track of earlier instructions or file contents.

The model is positioned as a fast and economical choice for agentic coding harnesses including Grok Build, Cursor, Hermes Agent, OpenClaw, Kilo Code, and OpenCode, and it also generalizes to other agentic and tool-calling workloads outside pure coding. xAI serves it at over 100 tokens per second, and independent reviews describe it as offering strong throughput at a low input and output price point. Stork AI's evaluation placed it in the middle of the field with a composite Intelligence Index score of 39.80, ranking twenty-seventh among one hundred fifty-five models, while noting that early public-beta users reported occasional hallucinations, a useful caution for teams deploying it on autonomous tasks.