PLANS / understand-grok-4-7-fast-limits

Plans

Understand Grok 4.7 Fast limits

Grok 4.7 Fast is the same Grok 4.7 model served on faster infrastructure at twice the standard token rates. Official: Grok 4.7, Pricing — Grok 4.7 Fast, Cursor Models. Announcement: Introducing Grok 4.7.

Where Fast runs

Surface Fast
Cursor Yes — billed through your Cursor plan
Grok Build (paid) Yes — at 2× standard token rates
Grok Build free tier Excluded
Public xAI API (api.x.ai) Unavailable — no Fast model id is published

The public models table lists grok-4.7 only. Do not invent a Fast API id.

xAI Fast rates (Build / Cursor docs)

Prompt tokens Input Cached input Output
Below 200k $4.00 / 1M $1.00 / 1M $12.00 / 1M
Above 200k $6.00 / 1M $1.50 / 1M $18.00 / 1M

Long-context rates apply once a request's prompt exceeds 200k tokens (Pricing).

Cursor long-context Fast multipliers

On Cursor's Cursor Models pool rows for Grok 4.7:

Condition Multiplier vs standard
Long context (>256k input), up to 500k
Fast mode
Fast mode + long context (>256k)

Published Cursor list prices for the four Grok 4.7 rows (input / cached / output per 1M):

Row Input Cached Output
Grok 4.7 $2 $0.5 $6
Grok 4.7 (Fast) $4 $1 $12
Grok 4.7 500k $4 $1 $12
Grok 4.7 500k (Fast) $6 $1.5 $18

Cursor plan gate for Fast

Start (India) includes non-fast Grok 4.7 at fixed medium effort. Fast mode and effort controls require Pro or higher (Start).

Pitfalls

  • Standard API calls use grok-4.7 — Fast is a Cursor / Build delivery option, billed at those surfaces' rates.
  • Cursor bills Fast through the Cursor plan; xAI's public API key path does not expose Fast.

Related