xAI Releases Grok 4.7, an AI Model Built for Coding and Work

At a glance

xAI has released Grok 4.7, its AI model for coding and professional work. The company says it can stay on long tasks for longer and check its own output more carefully. It is offered at the same price and speed as the previous model, Grok 4.6.

Key points

1. 1

   Grok 4.7 is trained to work on difficult tasks for longer and to re-check its finished output on its own.

2. 2

   The company says its ability to create documents and presentations is better than Grok 4.6's.

3. 3

   New safety measures refuse risky requests, such as those related to cybersecurity, while avoiding blocking legitimate security work.

4. 4

   Grok 4.7 can be used right away in Cursor and Grok Build, and is also available through the API. Pricing starts at $2 per million input tokens and $6 per million output tokens.

Why it matters

This news shows whether AI can take on long tasks and do so safely.