[xAI to Add Invisible Watermark to Grok API Text Output by Default Starting November 7th] According to xAI's official documentation, starting November 7th, it plans to add an invisible watermark to the API text output of Grok models such as 4.7 and 4.6 by default, covering plain text, structured output, and tool call parameters. The watermark is used to identify model-generated content and is applicable to global and regional API endpoints; team administrators can disable this feature by model.