CONFIRMED: xAI finally puts Grok 4.7 on the market

xAI announced Grok 4.7 on September 21, ending a cycle of promises, missed dates, and infrastructure clues that had kept the model in the realm of speculation. The release presents a system aimed at coding and knowledge work, with a larger base than Grok 4.6, a longer reinforcement-learning run, and more attention to tasks that can take hours. The central change is not simply a new version. It is xAI’s attempt to turn persistence and verification into a commercial advantage.

What changes for people using AI at work

According to xAI, Grok 4.7 was trained to sustain long contexts, check its own work more carefully, and handle problems that require multiple steps. The company also says it improves document and presentation creation, along with professional tasks measured by benchmarks such as GDPval and AA Briefcase. That matters more to teams that need to leave a conversation with a usable artifact than to users seeking only quick answers. For a small Brazilian business, the potential gain is a more continuous sequence from brief to research, analysis, and first deliverable, with fewer manual restarts.

Aggressive pricing, but do not confuse the headline with total cost

xAI lists a starting price of $2 per million input tokens and $6 per million output tokens. It also offers a fast variant with twice the output speed at twice the price. That is competitive for text-heavy agents and automations, especially against more expensive frontier models. The real cost still depends on context size, retry count, tools called, and the amount of output generated during verification. List price alone cannot tell whether a task becomes cheaper per completed delivery.

The important counterpoint

Independent evidence available before the announcement still reflected the product delay, while third-party catalogs showed inconsistent information, including prices far above the official statement. That does not invalidate the launch, but it argues for caution during integration. The public xAI documentation checked for this report still listed Grok 4.6 as the main model, a lag that must be resolved so developers can understand the identifier, limits, context, safety, and regional availability. The benchmarks cited are also selected by the company. They clarify the product thesis, but do not replace independent testing.

Our read

Grok 4.7 should enter the testing queue now for agents, coding, and document production, not production configuration by reflex. For small and midsize Brazilian companies, the most promising test is cost per verified delivery against models already in use, measuring persistence, review quality, latency, and rework in Brazilian Portuguese. xAI is making a clear promise: work longer without giving up too early. It now has to prove that promise survives real data, real tools, and tasks that do not fit a polished benchmark.

Sources

xAI, official Grok 4.7 announcement: https://x.ai/news/grok-4-7\nxAI, public model documentation: https://docs.x.ai/docs/models\nCometAPI, independent Grok 4.7 catalog: https://cometapi.com/models/xai/grok-4-7\nStartup Fortune, independent context on the earlier delay: https://startupfortune.com/elon-musk-shelves-grok-47-and-unveils-grok-48-instead