Text
Grok 4.7 — the new xAI model version in Twelver
Grok moves to 4.7. xAI prices 4.7 and 4.6 identically, so the new version comes at the old price.
What’s new
- 500K-token context, with both text and images accepted on input.
- Reasoning effort is selectable: low, medium, high, and the new xhigh.
- No rate change against 4.6.
- Worth knowing for long chats: above 200K prompt tokens a higher rate applies — to the whole request, not just the overage.
At a glance
| Detail | Value |
|---|---|
| Model | Grok 4.7 |
| Type | Text model; text and image input |
| Context | 500K tokens |
| Rate band | Higher above 200K prompt tokens |
| Where in Twelver | OpenAI-compatible API |
| API name | grok |
Frequently asked questions
How is Grok 4.7 different from 4.6?
A version up at the same price — xAI prices both identically. The 500K context is unchanged, and xhigh joins the reasoning levels.
Why does a long Grok conversation cost more?
Above 200K prompt tokens a higher rate applies, and it applies to the whole request rather than the overage. Since the conversation is re-sent in full on every turn, a long chat reaches that band on its own.
Do I need to change my code to get 4.7?
No. The API name is "grok" and it points at the current version, so requests with model: "grok" already run on 4.7. No separate xAI account required.
Try it
Open in Twelver chatRate your experience