The U.S. Office of Personnel Management (OPM) has removed Anthropic’s AI model Claude from its public inventory of artificial intelligence use cases and added Grok and Codex, according to an updated disclosure.

The removal of Claude follows a dispute between its developer, Anthropic, and the U.S. Department of Defense over AI guardrails. The conflict culminated in a government-wide ban issued by Donald Trump late last week, prompting several federal agencies, including OPM, to halt use of Anthropic’s AI models.

However, OPM said the addition of Grok and Codex was not directly tied to Claude’s removal. Agency spokesperson McLaurine Pinover told FedScoop that the tools were already being introduced as part of ongoing efforts to expand AI capabilities. According to the updated inventory, Grok is already in production, while Codex is currently being tested in a sandbox environment.

OPM also added several commercial AI-enabled systems to its disclosure, including Zendesk, Waze, Google Maps, and the Apple iPhone. The agency said it plans to update its reporting format and clarify risk classifications in future disclosures.