xAI · 2026-07-23 · major
Grok Build Workflows — xAI's coding CLI now fans a task across up to 1,024 parallel agents
Grok Build gained Workflows: orchestration scripts that fan a task across up to 1,024 parallel agents, verify results with independent skeptics, and return one report. Reusable slash commands live in .grok/workflows/.
Grok Build learned to fan a job across up to 1,024 parallel agents, verify with independent skeptics, and post one report — from a single slash command.
Key specs
| Default agents per run | 128 |
|---|---|
| Max agents per run | 1,024 |
| Model | grok-build-0.1 |
Quick facts
| Maker | xAI |
|---|---|
| Product | Workflows in Grok Build |
| Default parallel agents | 128 per run |
| Max parallel agents | 1,024 per run |
| Model | grok-build-0.1 |
| API price | $1 / 1M input tokens, $2 / 1M output tokens |
| Storage | .grok/workflows/ (reusable slash commands) |
| Availability | SuperGrok and X Premium+; heavy runs on SuperGrok Heavy |
What is it?
Workflows are xAI's new orchestration layer on top of Grok Build, the company's terminal coding agent. A workflow is a script that spawns many Grok Build sub-agents in parallel, gives each a slice of a larger job, and stitches the results back together. It targets work that will not fit in one context window.
How does it work?
Each workflow run defaults to 128 concurrent agents and scales up to 1,024. Sub-agent results pass through independent skeptic agents that try to refute the findings before they land in the final report, cutting false positives. Saved workflows live in .grok/workflows/ as reusable slash commands, so a team can commit them next to the code they operate on.
Why does it matter?
Fanning work across many agents is the current frontier of coding-agent design — Claude Code has fleet-view, Codex has multi-agent runs, and now Grok Build gets a first-class orchestration primitive. The 1,024-agent ceiling and the built-in skeptic verification make Grok Build competitive on the two jobs that break single-context coders: large PR reviews and 100-issue triage sweeps.
Who is it for?
Engineering teams using Grok Build for large PR reviews, backlog triage, or codebase-wide audits that a single chat cannot hold.
Frequently asked questions
- How many parallel agents can one Grok Build workflow run?
- A single Grok Build workflow runs 128 agents in parallel by default, and can scale up to 1,024 per run for the largest jobs. The orchestration script fans the task out, collects results, and pipes them through independent skeptic agents that challenge findings before the final report is written.
- What kinds of jobs are Grok Build workflows built for?
- xAI positions Grok Build workflows for tasks a single chat cannot hold in one context: reviewing every feature in a large pull request, triaging the last 100 open issues, auditing a codebase for one class of bug, and documentation sweeps. Anything that splits cleanly into independent pieces and needs a single consolidated report.
- How much does using Grok Build workflows cost?
- Access to workflows is included for SuperGrok and X Premium+ subscribers, but the largest parallel runs need SuperGrok Heavy, which sits around $300 per month. On the API, the underlying grok-build-0.1 model is billed at $1 per million input tokens and $2 per million output tokens; each workflow spends across every parallel agent it launches.
- Can a team save and share a Grok Build workflow across projects?
- Yes. A completed workflow can be saved as a reusable slash command in the .grok/workflows/ directory of any repo, checked into git alongside the code, and rerun by any teammate with a single command. That turns a one-off orchestration into a shareable playbook — the pattern lives with the codebase, not one engineer's chat history.
Try it
grok workflow run <name>