xAI Ships Grok 4.7 at the Same Token Price as Grok 4.6

xAI released Grok 4.7 on 21 September 2026 as its coding and knowledge-work model. Crypto Briefing describes a larger base model than Grok 4.6 and a longer reinforcement-learning run aimed at tasks that can take hours. The company says that training improved the model's ability to check its own work, hold a longer context, and operate inside the Grok Bot environment.

The published coding deltas are specific. CursorBench 4.0 moved from 40.4 percent on Grok 4.6 to 46.3 percent. Terminal Bench 4.0 moved from 20.3 percent to 38 percent. Those are the figures attached to the launch note, not an independent retest.

The price did not move

Grok 4.7 keeps the Grok 4.6 API rates: 2 dollars per million input tokens and 6 dollars per million output tokens. The report says the model is available through Cursor, Grok Build, and the Grok API, plus third-party coding tools, model routers, and cloud platforms. Grok 4.6 had arrived in August with the same long-running-agent pitch and the same price.

A separate write-up the same week attributed a 2.1 trillion parameter count to the model. xAI's own comparison pages have not always published that number, so the count is a report, not a spec sheet line. The parts that the launch note itself ties down are the benchmark deltas, the unchanged token price, and the distribution through existing coding tools.

Source: Crypto Briefing, 21 September 2026.