Elon Musk has announced the official launch of xAI's coding agent, Grok Build V1.0. The targets are clear: Anthropic's Claude Code and OpenAI's Codex. xAI is jumping straight into the AI coding agent market.
Musk's Grok Build: His Card in the Coding Agent Race
1. Plan mode: it shows an execution plan before changing code and allows review, edits, or a complete rewrite. It is a "plan first, execute later" workflow.
2. Parallel subagents: for changes to large codebases, it splits subtasks across multiple agents and runs them at the same time.
3. Terminal-first design: default vim keys, headless mode, and Agent Client Protocol (ACP) support let it connect to IDEs such as VS Code and JetBrains.
4. Extensibility: it supports a Claude Code-style ecosystem, including AGENTS.md, plugins, hooks, skills, and MCP servers.
Who Can Use It, and at What Price: Grok Build's Entry Point
The early beta was limited to the $300-a-month SuperGrok Heavy subscription, which drew criticism. It was later open-sourced, and the V1.0 launch widens access. Musk asked users for feedback, saying to tell him what to improve.
Why It's More Than Just Another Tool: The Spot xAI Is After
1. The coding agent market becomes a three-way race. After Claude Code and Codex comes xAI. The fiercer the competition, the more choices and lower prices developers get.
2. xAI's ecosystem strategy. Grok models, Cursor integration, and the Build tool. It is going after the whole environment developers work in, not just selling a model.
3. A caution. Early after launch, a security issue involving code repositories being uploaded to the cloud was reported. Check how your data is handled before adopting it.
Who Should Try It Now and Who Should Wait
If you already use Claude Code or Codex, there is no need to switch right away. But if you already use the xAI ecosystem (Grok models, Cursor) and are drawn to plan mode and parallel agents, it is worth testing. With a three-way race driving price and feature competition, everyone should get better within a few months.