Review
Grok 4.7 review: coding focus, API price, and SuperGrok plans
Key takeaways
- Grok 4.7 (Sep 21, 2026) is SpaceXAI’s flagship for coding and knowledge work, with a 500K context window.
- API starting point: $2 input / $6 output per 1M tokens under 200K prompt tokens; higher above that.
- Available via Grok API, Grok Build, Cursor, and consumer Grok apps on paid plans.
- Consumer ladder still centers on SuperGrok tiers (Lite / SuperGrok / Plus / Heavy).
- Product page: Grok. Branding may still say xAI in older docs; current launch posts say SpaceXAI.
Table of contents
What Grok 4.7 is
Grok used to mean “the funny X chatbot.” That is incomplete in 2026.
Grok 4.7 is SpaceXAI’s frontier model for longer coding and knowledge-work tasks, with real-time web and X search in the product surfaces and API access for builders.
What that means in practice:
- What it is: a coding-forward model with large context and live search gravity.
- What limits it: SuperGrok can cost more than ChatGPT Plus or Claude Pro; writing polish is still debated.
- What it is not: a free unlimited API, and not the same product as OpenAI Dots or Meta Muse.
SpaceXAI’s launch post pitches better long-horizon work, stronger self-checks, and the same list price/speed band as Grok 4.6 for the standard variant. Believe your evals over the chart in the blog.
Last verified: October 2026 against https://x.ai/news/grok-4-7, https://x.ai/pricing, and our Grok tool page.
API pricing
| Prompt size | Input / 1M | Cached input / 1M (release notes) | Output / 1M |
|---|---|---|---|
| Below ~200K tokens | $2 | ~$0.50 | $6 |
| Above ~200K tokens | $4 | ~$1 | $12 |
Also watch for:
- Fast variant: about 2× output speed at about 2× price where offered (Cursor / Grok Build callouts; confirm public API availability).
- Reasoning effort: low / medium / high / xhigh style controls in supporting clients.
If your prompts regularly cross 200K, model your bill with the higher tier before you celebrate the $2/$6 sticker.
Consumer SuperGrok plans
| Plan (commonly cited Oct 2026) | Price / month | Role |
|---|---|---|
| SuperGrok Lite | ~$10 | Lower entry, limited headroom |
| SuperGrok | ~$30 | Main individual paid tier |
| SuperGrok Plus | ~$100 | Higher usage, richer media limits |
| SuperGrok Heavy | ~$300 | Top usage / multi-agent style limits |
Exact inclusions (image/video, Expert mode, Grok Build access) move. Use the live pricing page as source of truth the day you publish or buy.
Coding vs general chat
| Job | Grok 4.7 fit | Often better elsewhere |
|---|---|---|
| Long coding agents | Strong launch focus | Claude Code / Sol if your harness already wins there |
| Real-time X + web context | Native strength | Perplexity-style research tools for citation-heavy briefs |
| Polished long-form writing | Mixed | Claude |
| Cheapest $20 generalist seat | Weak match | ChatGPT Plus / Claude Pro |
A simple way to think about it: pick Grok when live context and coding agents matter more than the cheapest general chat plan.
Who should use it
Good fit:
- Developers already in Cursor or Grok Build
- Teams that want API access around $2/$6 for coding workloads
- Users who live on X and need real-time social context
Skip or wait if:
- You need the lowest-cost general assistant at ~$20/mo
- Your org cannot standardize on SpaceXAI/xAI billing yet
- Writing quality on your docs corpus fails vs Claude in a blind test
Related directory page: Grok. For model neighbors in this cluster, see GPT-6.1 Sol and Claude.
FAQ
What is Grok 4.7?
Grok 4.7 is SpaceXAI’s (formerly xAI) September 2026 frontier model for coding and knowledge work. It offers a 500K-token context window and is available through the Grok API, Grok Build, Cursor, and consumer Grok apps on paid plans.
How much does Grok 4.7 cost on the API?
List pricing starts at about $2 per 1M input and $6 per 1M output below 200K prompt tokens, rising to about $4 / $12 above that. Cached input discounts apply per xAI release notes. A fast variant is sold at roughly 2x speed and 2x price where offered.
How much is SuperGrok?
Consumer SuperGrok plans are commonly cited at Lite $10, SuperGrok $30, Plus $100, and Heavy $300 per month. Confirm https://x.ai/pricing before you buy; plan names and inclusions change.
Is Grok 4.7 good for general writing?
It can write, but the launch focus is coding and multi-hour knowledge work plus real-time web/X context. Many teams still prefer Claude for polished long-form prose.
What is Grok Build?
Grok Build is SpaceXAI’s terminal coding agent surface for Grok models. Grok 4.7 is called out as available in Grok Build and Cursor in the launch notes.
1 curated tool below.
| Tool | Best for | Pricing | Billing note |
|---|---|---|---|
| Grok | AI chatbot / coding assistant | Freemium | Free Trial |
Frequently asked questions
What is Grok 4.7?
Grok 4.7 is SpaceXAI’s (formerly xAI) September 2026 frontier model for coding and knowledge work. It offers a 500K-token context window and is available through the Grok API, Grok Build, Cursor, and consumer Grok apps on paid plans.
How much does Grok 4.7 cost on the API?
List pricing starts at about $2 per 1M input and $6 per 1M output below 200K prompt tokens, rising to about $4 / $12 above that. Cached input discounts apply per xAI release notes. A fast variant is sold at roughly 2x speed and 2x price where offered.
How much is SuperGrok?
Consumer SuperGrok plans are commonly cited at Lite $10, SuperGrok $30, Plus $100, and Heavy $300 per month. Confirm https://x.ai/pricing before you buy; plan names and inclusions change.
Is Grok 4.7 good for general writing?
It can write, but the launch focus is coding and multi-hour knowledge work plus real-time web/X context. Many teams still prefer Claude for polished long-form prose.
What is Grok Build?
Grok Build is SpaceXAI’s terminal coding agent surface for Grok models. Grok 4.7 is called out as available in Grok Build and Cursor in the launch notes.