Anthropic’s Claude AI Breaks Free with Global Launch and Safety Upgrades

Anthropic just scored a major win. The US government lifted export restrictions on its latest AI models, Fable 5 and Mythos 5. This move unlocks global access and signals a new chapter in AI safety and collaboration. But what led to this dramatic shift? And why does it matter now?

Export Controls Spark a Race Against Time

On June 12, 2026, the US government slapped export controls on Anthropic’s Claude AI models Fable 5 and Mythos 5. The reason? Security concerns. Researchers at Amazon discovered a jailbreak that let Fable 5 generate exploit code. That was a red flag for national security. Suddenly, Anthropic’s cutting-edge models were off-limits internationally. Even US organizations lost access to Mythos 5 temporarily.

This action rattled markets too. CoinDesk reported that Anthropic’s pre-IPO perpetual contract dropped about 3.7% when access was suspended. The pressure was on Anthropic to fix these vulnerabilities — fast.

Anthropic Strikes Back with Bold Safety Measures

Anthropic didn’t just sit still. They launched a multi-pronged strategy to fight back against jailbreak exploits. They set up a program inviting hackers to red-team their models. That means ethical hackers actively tried to break the AI, exposing weak spots before bad actors could.

Anthropic also built a dedicated internal team to monitor jailbreak threats 24/7. This team tracks emerging risks and responds immediately. But that’s not all. They trained an improved safety classifier. This tool targets and blocks behaviors linked to jailbreaks. Requests flagged by this classifier are notified and rerouted to an earlier model, Opus 4.8, which handles them safely.

Anthropic proudly announced, “Over 99 percent of jailbreak attempts are currently blocked.” That’s a huge leap in AI security. Their CEO, Dario Amodei, spearheaded these efforts with full commitment to safer AI.

Government Collaboration Unlocks Global Access

The turnaround came fast. After intense collaboration with the US government, all export restrictions lifted by June 30, 2026. Access to Mythos 5 was restored to US organizations on June 26, and Fable 5 returned globally on July 1.

Howard Lutnick, US Commerce Secretary, sent a letter to Anthropic confirming the change. He said, “Anthropic has worked with the US government to address risks associated with the Covered Models.” He added, “Over the past two weeks, we have worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US Government and strengthen America’s leadership in AI.”

White House Chief of Staff Susie Wiles praised this partnership, saying, “The government and private sector have worked together in a way we have never seen before and this foundation of America First is unprecedented.”

Building a Safer AI Future Together

Anthropic isn’t stopping here. They’re teaming up with giants like Amazon, Microsoft, and Google to create a shared framework for assessing AI jailbreaks. This collaborative approach aims to set new security standards for frontier AI models worldwide.

The company plans to offer pre-release government access and evaluations for any AI models relevant to national security. They’ll also introduce rapid information sharing whenever jailbreaks or misuse patterns emerge. Dedicated teams will focus on government priorities, provide compute support, and share safety and red-teaming expertise.

• Global availability of Fable 5
• Mythos 5 access restored in the US
• Over 99% of jailbreak attempts blocked
• New safety classifier and rerouting systems
• Collaboration with top tech companies on safety standards
• Government partnerships for pre-release model reviews

This is a milestone moment. Anthropic showed that rapid innovation and tight security can go hand in hand. The US government demonstrated a new model for AI oversight — one based on partnership, not punishment. The future of AI safety just got a huge boost.

Will other AI labs follow Anthropic’s lead? The race to secure powerful AI is accelerating. But for now, Anthropic’s Claude models are back in the game, safer and more accessible than ever. The AI world is watching closely. And so are we.