Service update · xAI
xAI will start watermarking Grok text on its API by default, and team admins can opt out per model
- Effective time
- Review updated
No code changes are needed, and xAI says the watermark doesn't affect response quality or speed. It covers Grok 4.7, 4.6, 4.5, 4.3 and the grok-4.20 models in the Responses and Chat Completions APIs, including structured outputs and tool call arguments. xAI says it's doing this to comply with the EU AI Act. An x-watermark header shows whether a response was marked, and access to the detection API is by application. Opting out is set per model for the whole team, requires confirming that a legal exception applies, and some models may be locked on.