service xaivendor xAI (pages now branded “SpaceXSI”)category Model APIs
No (Favourable)Per the Enterprise Terms of Service (last updated Aug 14, 2026). De-identified aggregate data may still be used unless you use ZDR.
Per the Enterprise Terms of Service (last updated Aug 14, 2026). De-identified aggregate data may still be used unless you use ZDR.
SpaceXSI will not use any User Content to train any foundation models, large language models, or other artificial intelligence systems
Deleted within 30 days of the session ending, unless agreed otherwise, legally required, or needed for safety/abuse investigation.
All User Content will be automatically and permanently deleted no later than 30 days after the end of the interaction or session in which it was submitted
Against the 21 tools that state a period (grey dots):
Under ZDR, content is deleted within an hour and no logs are kept, including for safety.
no logs, backups, persistent copies, or other durable storage containing User Content will thereafter be retained for any purpose, including safety, debugging, or legal compliance.
Outside ZDR, retention can exceed 30 days for safety, moderation or abuse investigation.
reasonably necessary for safety, security, compliance, moderation, abuse prevention, or investigation of suspected violations of this Agreement
Global endpoint may route between regions. US endpoint (us.api.x.ai) pins handling to the US, costs 10% more, and supports fewer models.
SpaceXSI may route requests between regions for capacity or reliability, so the processing location is not guaranteed.
Report available to customers under NDA via the Trust Center. ISO 27001 not stated.
We are SOC 2 Type 2 compliant.
Terms require a BAA before submitting PHI, which implies one is offered.
unless Customer has first entered into a Business Associate Agreement or Amendment
The API is hosted only. xAI has published weights for older models (Grok 1, Grok 2) on Hugging Face; current API models are not downloadable.
This repository contains the weights of Grok 2, a model trained and used at xAI in 2024.
Spotted something wrong or out of date? Corrections are welcome; see how to send one. How statuses are assigned.