xAI · in Borg

Grok 4.7

Grok 4.7 is xAI's successor to 4.6, built on a larger base model with longer reinforcement-learning training for difficult, extended tasks.

Reasoning500,000 tokenstext, image

500,000 tokensContext window

1037 msMedian time to first token (gateway proxy)

33Output tokens per second (whole turn)

$2 / $6Vendor API price in / out · per 1M

Included in Borg

Download Borg Borg for Business

At a glance

Specification	Value
Provider	xAI
Family	Grok
Release date	2026-09-21 Vendor source
Context window	500,000 tokens Vendor source
Input modalities	text, image
Output modalities	text
Reasoning	Yes
Open weights	No
Knowledge cutoff	2026-05 Vendor source
Borg lane ids	`grok-4.7`, `x-ai/grok-4.7`
Versions in Borg	Grok 4.3 · Grok 4.5 · Grok 4.6 · Grok 4.7
Median time to first token (gateway proxy) on Borg	1037 ms
Median tokens per second (end to end) on Borg	33 output tokens per second

Not published by the vendor: thinking modes, fusion roles.

What Grok 4.7 is good for

Grok 4.7 is xAI's successor to 4.6, built on a larger base model with longer reinforcement-learning training for difficult, extended tasks. The vendor emphasizes better self-checking, sustained software engineering and professional knowledge work; it also reports improved document and presentation work. It supports configurable reasoning, text and image inputs, and text output, so it can work from a screenshot while writing code or explaining a document, but it is not a media-generation model. Unlike the cheaper 4.3, this is the flagship coding and general-purpose option; unlike 4.6, it brings a newer training run without changing the standard input and output token prices. Borg includes Grok 4.7 in the subscription, where Borg Fusion can route to it or you can choose its lane directly.

long-running agentic coding
complex terminal and engineering tasks
professional document drafting
self-checking technical analyses

How Borg uses it

Grok 4.7 is available as a subscription catalog lane in Borg. The catalog does not publish which Fusion profiles route to this model, so no routing role is claimed here. You can pick it directly by setting the OpenAI-compatible gateway model field to a listed lane id. Check the current catalog for availability before building a workflow around a lane.

model: "grok-4.7"

Measured on Borg

Metric	Median	95th percentile	Samples
Time to first token (gateway proxy)	1037 ms	4865 ms	1436
Tokens per second (end to end)	33 output tokens per second	89.8 output tokens per second	1467

Time to first token (gateway proxy)

Lower is fasterFleet median 3971.5 ms

Grok 4.71037 ms

Output tokens per second (whole turn)

Higher is fasterFleet median 31.9 tokens/s

Grok 4.733 tokens/s

Time to first token (gateway proxy): time until the first streamed response event reaches the gateway; our closest available measure of time to first token
Tokens per second (end to end): tokens per second, end to end: provider-reported output tokens divided by the whole turn's elapsed time, including queueing and reasoning; not a streaming per-token rate

Completion speed covers the whole turn including queueing and reasoning time; it is not a streaming tokens-per-second figure.

Measured on Borg: SearchAtlas AI Gateway production request logs. 30-day window ending 2026-10-08; 1681 requests. Minimum 100 samples per published request metric.

Benchmarks

Benchmark	Vendor-reported score	Source
CursorBench 4.0	46.3%	Vendor source
DeepSWE v1.1 (high effort)	71.0%	Vendor source
EEBench	64.0%	Vendor source
Terminal-Bench 4.0	37.6%	Vendor source
HealthBench Professional	56.7%	Vendor source

Pricing

Included in Borg subscription. Vendor API list price: $2 input / $6 output per 1M tokens. Vendor price source.

Global rates for prompts below 200,000 input tokens. At 200,000 input tokens or more, all tokens use long-context rates: $4 input / $12 output / $1 cached input per 1M tokens. US regional inference costs 1.1× the global rate. Vendor pricing terms

Compare subscription and API pricing · Cost Savings Calculator

FAQ

What is Grok 4.7?

Grok 4.7 is listed in Borg's subscription catalog as grok-4.7. Grok 4.7 is xAI's successor to 4.6, built on a larger base model with longer reinforcement-learning training for difficult, extended tasks.

Is Grok 4.7 in Borg?

Yes. Grok 4.7 is listed in Borg's subscription catalog under the lane grok-4.7. You can choose the lane directly, or explore the available Fusion profiles. Availability and model details can change; check the current catalog before deployment.

What context window does Grok 4.7 have?

The published context window is 500,000 tokens. Consult the cited sources for details and current limits. Vendor source

Is Grok 4.7 good for coding?

Vendor-published evaluations appear in the benchmarks section; their settings may differ from your tasks. Its listed uses include long-running agentic coding, complex terminal and engineering tasks, professional document drafting, self-checking technical analyses. Try the lane for your own tasks, and compare the output with your requirements before choosing a default.

How fast is Grok 4.7 on Borg?