xAI
xAI releases Grok 4.7 at Grok 4.6's unchanged $2/$6 pricing
Beats Grok 4.6 on all seven of xAI's benchmarks. Independent testing says it needs about twice the tokens to finish a task.
The answer
Grok 4.7 launched 21 September 2026; xAI kept Grok 4.6's $2/$6 per-million-token price.
xAI released Grok 4.7 on 21 September 2026. It is live in the Grok API, Cursor and Grok Build, and GitHub began a phased rollout across Copilot Pro, Pro+, Max, Business and Enterprise the same day.
Grok 4.7 uses a new, larger base model compared to Grok 4.6.
Pricing and availability
Pricing is unchanged from Grok 4.6: $2 per million input tokens and $6 per million output for prompts under 200,000 tokens, doubling to $4/$12 above that. A fast variant, limited to Cursor and Grok Build, costs twice the standard rate. Azure AI Foundry and AWS Bedrock do not yet offer it; both top out at Grok 4.3.
| Model | Input / output per M tokens |
|---|---|
| Grok 4.7 | $2 / $6 |
| GPT-6 Sol | $2 / $10 |
| Claude Opus 5.5 | $4 / $20 |
| Claude Fable 5.1 | $10 / $50 |
Benchmarks
On xAI's own seven-benchmark table, Grok 4.7 beats Grok 4.6 across the board. Claude Fable 5.1 Max still leads on CursorBench 4.0, Terminal-Bench 4.0, GDPval and HealthBench Professional; Grok 4.7 leads Fable on EEBench and the Harvey Legal Agent Benchmark. Artificial Analysis scored Grok 4.7 at 46 on its Intelligence Index v4.3.2, against Grok 4.6's 44 on the same version — not comparable to the 61 Grok 4.6 was quoted at on an earlier index version.
Grok 4.7 (xhigh) uses approximately 81k output tokens per Intelligence Index task, compared with 36k for Grok 4.6 (high) and 27k for GPT-6 Astra (max)
Safety, delays and what's next
xAI says a new safeguard stack scored 62.4% on LatchBio's biosafety benchmark and blocked all but 3.3% of dangerous dual-use prompts on its internal HackerBench v0.3 evaluation. The model had slipped repeatedly since late July. Musk has said Grok 4.8, then Grok 4.9, then Grok 5 will follow, with no dates given.
A day after Grok 4.7 shipped, Anthropic released Claude Opus 5.5 at $4/$20 per million tokens and OpenAI released GPT-6 Sol at $2/$10 and GPT-6 Luna at $0.10/$0.50 — both cuts on their prior generations. xAI was the only one of the three that did not move on price.
Coverage has widely cited a 2.1-trillion-parameter figure for Grok 4.7's base model and supplementary SpaceX training data; neither appears in xAI's launch materials, and both trace to Elon Musk's own posts.
Frequently asked questions
What does Grok 4.7 cost?
Is Grok 4.7 available in GitHub Copilot?
How does it score against Grok 4.6?
Does it cost more to run than Grok 4.6 despite the same price?
Sources
- Introducing Grok 4.7 — xAI, 21 September 2026
- Grok 4.7 is now available in GitHub Copilot — GitHub, 21 September 2026
- Grok 4.7 scores 46 on the Artificial Analysis Intelligence Index to bring SpaceXAI into the top 4 AI labs — Artificial Analysis, 21 September 2026
- Grok 4.7 pairs coding gains with the same affordable pricing — but high token consumption threatens real-world ROI — VentureBeat, 21 September 2026
- Grok 4.7 Brings Big Coding Upgrades to Challenge Claude AI at Unchanged Pricing — Android Headlines, 21 September 2026
- Grok 4.7 pricing: API token costs, tiers, and the Fast tax — eesel AI, 23 September 2026
- xAI launches Grok 4.7 at $2 per million input tokens — tbreak, 22 September 2026