Commerce clears Mythos 5 for ~100 US organizations; Fable 5 remains blocked
Commerce Secretary Howard Lutnick cleared Anthropic to restore Mythos 5 — its cybersecurity-focused model — to roughly 100 approved organizations and federal agencies, ending two weeks of total global access blackout after a June 12 national-security order. Fable 5 stays offline, and the episode has established a visible precedent: the US government can and will halt frontier model deployments, a pattern that also delayed OpenAI's GPT-5.6 until government-approved partners were cleared.
Source: fortune.com ↗
Anthropic has worked with the US government to address risks associated with the Covered Models. These efforts have yielded significant progress.
Commerce Secretary Howard Lutnick
Why this matters
- → US government now directly controls frontier AI model deployments via security vetting.
- → Sets precedent for conditional access that other providers (OpenAI) are already following.
- → Exposes gap between theoretical safeguards and government confidence in jailbreak resistance.
Gatekeeping AI release