Grok Build (xAI) | Ry Walker Research

Key takeaways

Officially shipped: xAI launched Grok Build in early beta on May 14, 2026 — a Rust-based, terminal-native coding agent installed via curl, initially gated to SuperGrok Heavy ($300/mo), then opened to SuperGrok and X Premium Plus subscribers
Powered by grok-build-0.1, a coding-specific model with 256K context at $1.00/$2.00 per million input/output tokens via API — it replaces the now-deprecated grok-code-fast-1
Shipped features are conventional, not exotic — Plan Mode with reviewable diffs, parallel subagents in isolated git worktrees, headless CI mode, MCP/AGENTS.md/plugins support. Pre-launch claims of "Arena Mode" and local-first privacy did not appear in the official product
Steepest entry price in the category — full access launched at roughly 15x the cost of Claude Code or Codex CLI bundled subscriptions, and no independent SWE-Bench Verified score for grok-build-0.1 was published at launch

FAQ

What is Grok Build?

xAI's terminal-native coding agent, launched in early beta on May 14, 2026. A Rust CLI powered by the grok-build-0.1 model (256K context) featuring Plan Mode, parallel subagents running in isolated git worktrees, and a headless mode for CI. Available to SuperGrok and X Premium Plus subscribers.

How does Grok Build compare to Claude Code?

Feature-wise it closely mirrors Claude Code — plan-then-execute workflow, subagents, MCP, AGENTS.md, hooks, headless CI. It differs on price (full access launched via SuperGrok Heavy at ~$300/mo versus Claude Code bundled at $20/mo), context window (256K vs 1M), and maturity (early beta, no published SWE-Bench Verified score for its model at launch).

Overview

Status note (June 2026): Grok Build is shipping — xAI launched it in early beta on May 14, 2026. But the product that shipped differs materially from the pre-launch coverage this profile was originally based on. Earlier third-party reports of "8 parallel agents," an algorithmic "Arena Mode," local-first privacy, and npm installation did not appear in the official launch; independent analysis flagged much of that coverage as overstated. This profile has been rewritten against the shipped product.

Grok Build is xAI's terminal-native coding agent — a CLI written in Rust, installed via curl -fsSL https://x.ai/cli/install.sh | bash, that plans, edits, and runs code from the terminal with diff-based review. It is xAI's answer to Claude Code and Codex CLI, arriving roughly a year behind both.

Key stats (as of June 2026): grok-build-0.1 model, 256K token context, $1.00/$2.00 per million input/output tokens via API, early beta.

What Shipped vs. the Rumors

Pre-launch coverage claimed	What actually shipped
Up to 8 parallel agents, side by side	Parallel specialized subagents (research, implementation, review) in isolated git worktrees — no official agent count
"Arena Mode" algorithmic ranking of agent outputs	Not present in official launch materials
Local-first; code never leaves the machine	Standard cloud-model agent requiring xAI account sign-in
`npm install -g grok-build`	curl install script from x.ai
Powered by grok-code-fast-1 (70.8% SWE-Bench Verified)	Powered by grok-build-0.1; grok-code-fast-1 is deprecated and no longer listed in xAI's model docs

Features

Plan Mode

For complex tasks, Grok Build starts in plan mode: it drafts a structured, step-by-step approach that the developer can approve, comment on per-step, or rewrite entirely before execution begins. Once approved, every change surfaces as a clean reviewable diff — the same review pattern Claude Code and Codex established.

Subagents and Worktrees

Larger tasks are delegated to specialized subagents that run in parallel — research, implementation, and review concurrently. Subagents launch in their own isolated git worktrees, preventing conflicting edits on the main branch.

Headless CI Mode

grok -p "task" runs non-interactively with optional streaming-json output. Default permission mode is ask (prompt per tool call); --always-approve enables fully automated pipelines.

Extensibility

MCP servers, AGENTS.md conventions, plugins, skills, and hooks are supported, with grok inspect exposing configuration sources — a near-exact mirror of the Claude Code extension surface.

Model and Pricing

Item	Detail
Model	grok-build-0.1 (coding-specific, text + image input)
Context window	256K tokens
API price	$1.00 / $2.00 per million input/output tokens
Subscription access	Launched on SuperGrok Heavy ($300/mo); expanded to SuperGrok and X Premium Plus during early beta
Flagship lineup context	xAI's current flagship is Grok 4.3 (1M context, $1.25/$2.50 per M tokens)
Surfaces	Terminal only (TUI + headless CLI); no IDE or GitHub-native integration at launch

Viability and Adoption

Grok Build is two years into xAI's pattern of uneven developer-tool follow-through, and the launch is real but narrow. The product is explicitly labeled early beta; no IDE surface exists; and xAI published no SWE-Bench Verified score for grok-build-0.1 at launch. The older 70.8% figure belonged to the now-deprecated grok-code-fast-1 and trails the June 2026 leaderboard (Claude Opus 4.7 Adaptive at 87.6%, OpenAI Codex at 85%) by roughly 17 points. No verified user or revenue numbers for Grok Build have been published as of June 2026.

Cautions

Pricing is the category's steepest. Full access launched behind SuperGrok Heavy at roughly $300/month versus Claude Code or Codex CLI bundled into $20/month plans — independent reviewers called out the ~15x cost differential as the product's biggest adoption barrier.
Benchmark claims rest on xAI's own evaluation suite, without independent validation at launch.
The hype gap is documented. Critical coverage found launch-cycle claims about context window, Plan Mode defaults, and the model's identity to be overstated or wrong — a caution for anyone evaluating the tool from secondhand coverage.
Early beta, smallest track record. Every competing foundation-lab agent has 12+ months of production hardening; Grok Build has weeks.

Competitive Position

Strengths: Genuine parallel subagent architecture with worktree isolation. Cheapest coding-model API pricing in the lineup ($1/$2 per M tokens). Full extension stack (MCP, AGENTS.md, hooks) at launch rather than added later.

Weaknesses: Most expensive subscription gate in the category. Smallest context (256K vs 1M for Claude Code and Gemini CLI). Terminal-only — no IDE, no GitHub integration. Early beta with no published independent benchmarks. xAI model lock-in.

Bottom Line

Recommended for: Existing SuperGrok / X Premium Plus subscribers who want a Claude Code-style agent within the xAI ecosystem; teams already paying for SuperGrok Heavy that want parallel subagent workflows with worktree isolation.

Not recommended for: Anyone optimizing for cost (15x the entry price of Claude Code/Codex), teams needing IDE or GitHub-native surfaces, or buyers who require independently benchmarked models before adoption.

Outlook: The May 2026 launch proves xAI will ship developer tools, but the offering is a fast follow rather than the differentiated product pre-launch coverage promised — Arena Mode and local-first privacy never materialized. Watch whether xAI publishes independent benchmarks for grok-build-0.1, drops the subscription gate toward a $20-30 tier, and sustains release cadence past the beta; xAI's developer-tool follow-through has historically been uneven, and the deprecation of grok-code-fast-1 within months is a reminder of churn risk.

Research by Ry Walker Research

Sources