At a Glance
| Vendor | xAI |
| Family | Grok 4 series |
| Launched | xAI released Grok 4.7 on September 21, 2026, about six weeks after Grok 4.6. xAI, which has operated as SpaceXAI since a July 2026 rebrand, describes it as "the most powerful model for coding and knowledge work. Twice as fast, at half the price of comparable models." It says the model is built on a larger base than 4.6, with more reinforcement learning on long multi-hour tasks and stronger self-checking. |
| Context window | 500,000 tokens, unchanged from Grok 4.5 and 4.6. Input is text and image. Launch coverage gives a May 2026 knowledge cutoff. |
| Pricing | $2 per million input tokens, $0.50 per million cached input tokens, and $6 per million output tokens for prompts under 200,000 tokens, the same as Grok 4.6. From 200,000 tokens the whole request is billed at $4 / $1 / $12. Grok 4.7 Fast runs at about twice the output speed for twice the price. |
| Access channels | xAI API (grok-4.7), Grok Build, Cursor, GitHub Copilot, Vercel AI Gateway, and other third-party routers and cloud platforms. xAI also says the model understands the Grok Bot harness natively, which is what Team Bots run on. |
Notable Benchmarks
xAI's launch table shows gains over Grok 4.6 on all seven benchmarks it lists, the largest on Terminal-Bench 4.0 at 38.0 percent against 20.3 percent. xAI says it leads GPT-5.6 Sol at max effort on four of the seven and trails Claude Fable 5.1 on CursorBench, Terminal-Bench, GDPval, and HealthBench. Independently, Artificial Analysis gives it an Intelligence Index score of 46, ranking 16th of 655 models, at about 188 tokens per second, and notes it uses about 81,000 output tokens per task against 36,000 for Grok 4.6, which raises real cost above what the unchanged list price suggests.
Strengths
Fast serving, low output pricing, a large step up from 4.6 on terminal and agentic coding tasks, and tight integration with xAI's own agent products. The live X data surface remains unique among frontier labs.
Limitations
Still behind the leading frontier models on independent composite indices. Higher token use per task offsets part of the price advantage. The launch material describes no multimodal improvements and no change to live search or X search behaviour, and the context window is half that of the GPT-6 and Claude 5 families.
Brand-Visibility Implications
Grok 4.7 changes little about how Grok finds information and a fair amount about where Grok is used. Retrieval still leans on X, so the brands that do well in Grok answers are the ones with an active, well-engaged presence there. What is new is distribution into work: 4.7 is the model xAI ties to Grok Build and to Team Bots, the shared agents it launched a week later for Slack. That puts Grok into vendor discussions inside companies, not only into consumer questions on X. A May 2026 cutoff also means coverage from early 2026 is now in recall without retrieval. Track Grok separately from other assistants. See Grok Team Bots, Grok citation patterns, X Original Content Rewards, and Grok brand visibility.
How Presenc AI Tracks This Model
Presenc AI monitors brand visibility on xAI's Grok 4 series as part of continuous multi-platform AI visibility tracking. We sample xAI Grok 4.7 across representative prompt sets daily, compare against competitor performance on the same prompts, and flag material mention-rate changes so brand teams can respond quickly when AI representation shifts.