Breaking

SpaceXAI Adds Grok 4.7 Fast to Grok Build and Cursor with a 2x Output-Speed Tier

September 22, 2026 at 04:16 EDT

Foundation Models
AI Agents
Software Dev & Coding

SpaceXAI has made Grok 4.7 Fast available in Grok Build and Cursor, serving the same Grok 4.7 model on faster infrastructure with a stated 2x output-speed tier and premium token pricing. The Fast option is excluded from Grok Build's free tier and is not available through the public xAI API, while standard Grok 4.7 is now available as grok-4.7.

Sep 21, 2026 · SpaceXAI

Grok 4.7 Fast trades speed for cost

The same Grok 4.7 model moves onto faster serving infrastructure, pairing a stated 2x output-speed tier with higher token prices for coding-agent work.

stated output speed

$4/M

Fast input, below 200k

$12/M

Fast output, below 200k

500k

maximum context

The multiplier is literal

standard baseline

Fast speed tier

headline rate tier

The product pitch is not a new model: it is faster serving for the same model at a premium tier.

Terminal-Bench 4.0 nearly doubles

20.3%

Grok 4.6

38.0%

Grok 4.7

The largest benchmark gain is on multi-hour terminal work, where Grok 4.7 moves from 20.3% to 38.0%.

Why builders care

A faster stream can help long codebase edits, debugging runs and agent sessions where waiting on output slows review and iteration.

Why teams will test

Higher token rates and verbose reasoning can erase savings unless the model finishes real tasks with fewer retries and less human repair.

Production test

Track cost per completed task, latency, retries and human intervention. Token price and output speed are inputs, not the final bill.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.