Composer 2.5 enters Grok Build as xAI’s long-task coding model

xAI is pushing its coding model deeper into the Grok Build workflow. In a June 1 post, the company said Composer 2.5 is now available inside Grok Build.

The meaningful product claim is that Composer 2.5 is built for “long-running tasks” and complex instructions. That points beyond autocomplete toward coding-agent work: planning edits, following multi-step requirements, navigating project context, and carrying changes through execution. The source tweet says Composer 2.5 is available in Grok Build; a follow-up points users to x.ai/cli and the /models menu.

xAI’s Composer 2.5 page is indexed with a June 1 date and labels Grok Build as the current home for xAI’s latest coding model. Direct access to x.ai was blocked by Cloudflare from this crawl environment, so this article avoids adding unsupported benchmark or pricing claims. The verified engagement from FxTwitter is still material: more than 4,000 likes, roughly 480 reposts, and more than 640,000 views in the first day.

What to watch next is whether the long-task claim holds in real developer workflows. The useful tests are context retention, tool-call reliability, repository search, test execution, and the quality of explanations after changes land. For xAI, Composer 2.5 is also a distribution test: Grok Build has to feel useful in CLI and model-selection flows, not only as another model name in a menu.