✳ the wire · analysis

Grok 4.8 shows up in xAI's own code: its public coding agent already lists it

Grok 4.8strongby ArtificialWatch

Get launch alerts like this, free →

Grok 4.8 shows up in xAI's own code: its public coding agent already lists it
Source imagery · multiple independent credible sources

xAI's public grok-build repository, the source of its Grok Build coding agent, now lists grok-4.8 by name. The model picker registers "grok-4.8" as "Grok 4.8", and the router tests route to it with 500K and 256K context windows. We read the code on GitHub on September 29; the repo was last updated that evening.

WHAT IS SOURCED

The id and the name are in xAI's own code, which is stronger than a screenshot or a leaker's word. The context windows are test fixtures, not a spec sheet, so treat them as what xAI's tests assume rather than what the model will ship with. A separate report, from a single account, says Cursor's server-side model list already carries grok-4.8-high, grok-4.8-high-fast and grok-4.8-xhigh-fast; we have not seen that list ourselves.

WHAT WE DO NOT KNOW

No date, no price and no benchmark. xAI has not announced it; the only public word before this was Elon Musk describing a roughly 2.5T-parameter model on a new C++ stack. Grok 4.7 shipped on September 21, and xAI's recent cadence has been five to six weeks.

Grok 4.8 is armed on our radar, and every id shape above fires it.

Source: xai-org/grok-build on GitHub, read directly · a Cursor model-list report via @LuminaBench (single source) ↗ · Grok 4.8 tracker · the bench index

sweeping every 60 seconds

Know the minute it drops — not the minute we write it up.

Claude Opus 5 went live at 16:51 UTC. The alert was in subscribers' inboxes at 16:52.

  • Free forever
  • No card
  • Unsubscribe in one click

← back to the wire