xAI released Grok Build on May 14, 2026 — an agentic coding tool and command-line interface aimed at professional software engineering. The beta is open to SuperGrok Heavy subscribers and runs on Grok 4.3 beta, built on the same 16-agent “Heavy” architecture xAI introduced earlier this year.

What it does

Grok Build runs up to eight agents in parallel and carries a 2-million-token context window. For non-trivial work it opens in plan mode: it drafts an execution plan you can approve, comment on step by step, or rewrite entirely before any code is written. Once a plan is approved, every change lands as a reviewable diff rather than a direct write to your branch — the same approve-then-execute loop Claude Code and Cursor adopted earlier this year.

Pricing and access

Grok Build is bundled with SuperGrok Heavy, normally around $300/month. xAI is running an introductory rate of $99/month for the first six months, a 67% discount aimed at early adopters. There is no free tier and no standalone plan during the beta — access is gated behind the top SuperGrok subscription, which puts day-one evaluation out of reach for anyone not already paying for it.

Where it fits

Grok Build is the fourth major coding agent to ship a terminal-first CLI, after Anthropic’s Claude Code, OpenAI’s Codex, and Google’s Gemini CLI. Its plan-first, multi-agent shape closely mirrors what Claude Code’s Agent View and Cursor’s Composer 2 already do; the differentiation xAI is betting on is model quality and context size, not workflow novelty.

Whether a fourth entrant moves developers who already run a multi-tool stack is the open question, taken up in The coding-agent race gets crowded.