xAI entered the coding-agent race with Grok Build — a terminal-based agent CLI taking direct aim at Claude Code and OpenAI Codex. Currently early beta, available to SuperGrok Heavy subscribers at $300/month.
## What it does
Runs from the terminal, driven by natural-language prompts. It generates implementation plans, edits files, executes shell commands, manages dependencies, and supports collaborative workflows through subagents and plugins. The standout capability: it spawns up to 8 concurrent AI agents that simultaneously plan, search documentation, and write code.
## The model underneath
Grok Build runs on Grok 4.3 beta, which uses a 16-agent “Heavy” architecture and a 2 million token context window. There’s also a “plan mode” — the agent presents a logical plan you review, edit, and approve before it touches the codebase. That approval gate is the same governance pattern showing up across team-focused coding agents.
## Why it matters
The coding-agent field is genuinely crowded now: Claude Code, Cursor, Codex, Gemini’s Antigravity, and now Grok Build. xAI’s differentiator is raw parallelism (8 concurrent agents) plus the 2M context window. At $300/month it’s aimed squarely at power users — a bet that serious developers will pay frontier prices for frontier throughput.
Top AI Product 
Leave a comment