xAI's fast agentic coding model — the engine behind the Grok Build CLI, built to write and ship software.
Last verified · 2026-06-25 · by Moe Ameen
Grok Build 0.1 is a coding model from xAI, trained specifically for agentic software engineering — writing, editing, and debugging code across a whole project rather than answering one-off questions. It is the model that powers the Grok Build CLI, xAI's terminal coding agent, and xAI opened it to developers directly through the xAI API in public beta in late May 2026 (the model first reached early access earlier that month). The API model string is `grok-build-0.1`. It is a proprietary, closed-weight model; xAI has not published a parameter count.
It is positioned as a fast, reasoning-style coding model rather than a frontier generalist. On Artificial Analysis's independent benchmark of the June 2026 ("0616") snapshot it posts an Intelligence Index of 40 — above the median for models in its price tier — while running at roughly 104 output tokens per second with a time-to-first-token near half a second, which is quick for a model that does chain-of-thought work. It carries a 256k-token context window and accepts text and image input, returning text output. API pricing is $1.00 per million input tokens and $2.00 per million output tokens, with cached input billed at $0.20 per million (an 80% discount).
The Grok Build CLI around it leans on agentic patterns: plan-first execution (it proposes a plan before touching files), parallel sub-agents that can each work in their own Git worktree, and native Model Context Protocol (MCP) support so it can call external tools. That puts it in the same lane as Anthropic's Claude Code and OpenAI's Codex CLI — terminal-native coding agents rather than chat windows.
One thing to be clear about: Grok Build 0.1 is a developer tool. It writes software; it does not generate images, video, audio, or social posts, and it does not publish anything. If you found it while looking for a way to make content, it is the wrong category — but it is a strong way to build the thing you then make content about.
Grok Build 0.1 builds the product. Kompozy markets it. That is the clean division of labor, because the model writes code and nothing else — it does not make a launch video, a feature carousel, a changelog thread, or a single published post. The most common real workflow for a founder or indie builder is: ship a feature with Grok Build in the morning, then need a week of content announcing it by the afternoon. Drop the changelog, the README, or a two-line description of what you shipped into Kompozy as a source, and it fans that one input into a demo-style short, a carousel walking through the feature, a launch thread for X, a LinkedIn post, and a blog article — all in your brand voice through a Persona Brief — then schedules and publishes them across all nine connected platforms plus email and blog from one queue.
There is also a literal wiring angle, since Grok Build is an agent that writes integrations. It has native MCP support and is built to produce webhooks and automations, and Kompozy accepts a Custom Webhook as a blog/publishing destination. So you can have Grok Build stand up the glue — a webhook that pushes your release notes into your pipeline, or a script that drops new product shots into a folder Kompozy watches — and let Kompozy turn that raw material into finished, scheduled posts. Grok Build owns the engineering; Kompozy owns the audience.
Grok Build 0.1 is xAI's coding model trained for agentic software engineering — writing, editing, and debugging code across a project. It is the model behind the Grok Build CLI and is available to developers via the xAI API (model string grok-build-0.1). It opened in public beta in late May 2026.
Via the xAI API, grok-build-0.1 is priced at $1.00 per million input tokens and $2.00 per million output tokens, with cached input at $0.20 per million (an 80% discount). It has a 256k-token context window. Using it inside the Grok Build CLI may also be available through xAI subscription tiers.
No. It is a coding model — it writes and ships software and produces no images, video, audio, or social posts. To turn what you build into published content, you pair it with a content engine like Kompozy that generates the media and publishes across platforms.
All three are terminal-native agentic coding tools. Grok Build leans on plan-first execution, parallel sub-agents in Git worktrees, and native MCP support, and is positioned as a fast, lower-cost coding model. Which one fits depends on your stack and benchmarks; treat head-to-head claims as snapshots since all three ship quickly.
Build the product itself — a landing page, an app, an internal tool, or custom automations and webhooks. Then use Kompozy to generate and publish the marketing content around what you shipped: launch shorts, carousels, threads, blogs, and newsletters across nine platforms.