After a rollercoaster negotiation process with the Trump administration that dragged on for two weeks, Anthropic’s Mythos 5 is finally back in action — at least, somewhat, for a select group of organizations, according to a letter from the government to Anthropic that was viewed by The Verge. Fable 5, however — the public-facing Mythos-class model — appears to still be in limbo, with no apparent timeline for a rollout agreement.
The letter, dated June 26th and sent by Commerce Secretary Howard Lutnick to Anthropic co-founder Tom Brown, who had been recently leading negotiations, states that there’s been a “revision to the license requirements” based on the fact that Anthropic has recently “worked with the U.S. government to address risks” associated with Mythos 5 and Fable 5.
Anthropic spokesperson Danielle Ghiglieri told The Verge in a statement that the company had “received notice from the US government that Mythos 5, our strongest cybersecurity model, can be redeployed to a small group of cyber defenders and infrastructure providers.” She added, “We are working to provision the approved set of providers and restore their access to Mythos 5 as quickly as possible. We are pleased to see this progress and continue to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again.”
The US government did not lift the export control directive it hit Anthropic with two weeks ago, which barred any foreign national from accessing either model (including Anthropic’s own employees). Instead, the government essentially made an exception for Mythos 5, approving a select group of organizations for access the same way it did for OpenAI’s GPT-5.6, announced earlier today. Under this exception, Anthropic employees who aren’t US nationals and members of the approved organizations who aren’t US nationals are all greenlit to access Mythos 5, according to the letter.
“These efforts have yielded significant progress,” Lutnick wrote. “In addition, Anthropic has committed to work with the U.S. government on protocols and standards and releases for [Mythos-class models]. In light of this progress ... I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model.”
Pressure has been building on the Trump administration to make a change in the case-by-case type of regulatory environment it recently adopted, especially as Anthropic’s competitors’ cybersecurity-focused models kept getting better and better — and even pulling ahead of it on some cybersecurity-focused benchmarks. Pressure was also building within the US AI industry, particularly fears about the AI advancements China could make while top US AI labs were sidelined. Plus, there was the fact that top US government departments, like the National Security Agency, had lost access to Mythos 5.
Now, Anthropic has essentially the same deal OpenAI has: Release the model in a limited preview setting, for approved organizations only, such as trusted enterprises and the US government itself. Both AI labs hope that more general availability comes soon, both for enterprise deals and for public access (like with Anthropic’s Fable 5), but it’ll be ultimately up to the Trump administration. Though many have called for AI regulation, including AI labs themselves, a current consensus among some tech leaders is that the past two weeks haven’t been the way to do it.
In fact, OpenAI wrote in the company’s GPT-5.6 blog post, “We don’t believe this kind of government access process should become the long-term default. It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them. We are taking this short-term step because we believe it is the strongest path to broader availability in the coming weeks, while we work with the Administration to develop the cyber Executive Order framework and a repeatable process for future model releases.”
Lutnick wrote in his letter to Anthropic, “All other requirements of the June 12 letter remain in effect until further notice.” He added, “I reserve the right to reevaluate and adjust the scope of license requirements on [Mythos 5 and Fable 5], should circumstances change.”
Background on Anthropic and the Mythos Class
Anthropic, founded in 2021 by former OpenAI employees including Dario and Daniela Amodei, has positioned itself as a safety-first AI company. Its flagship model family, Claude, is known for its emphasis on ethical behavior and alignment. The Mythos class represents Anthropic's most advanced systems, designed specifically for cybersecurity applications. Mythos 5, released earlier this year, was hailed as a breakthrough in AI-driven threat detection and automated incident response. It was being used by several Fortune 500 companies and critical infrastructure operators before the government intervention.
The underlying technology of Mythos 5 integrates a novel architecture that combines transformer-based language models with specialized cyber reasoning modules. This allows the system to analyze network traffic, identify zero-day vulnerabilities, and recommend mitigation strategies in real time. Early benchmarks showed Mythos 5 outperforming existing tools in penetration testing and vulnerability assessment tasks by significant margins.
However, the Trump administration's June 12 directive blindsided Anthropic. The export control order cited national security concerns, stating that the models could be used to develop offensive cyber weapons or be exploited by foreign adversaries. The directive effectively froze all non-US access, including to Anthropic's own research teams, many of whom are based in Canada and the UK.
The Negotiation Process
The two weeks that followed were marked by intense behind-the-scenes lobbying. Anthropic CEO Dario Amodei personally met with Commerce Secretary Lutnick and national security officials. The company submitted extensive documentation demonstrating the model's safety features, including its refusal to generate harmful code and its built-in usage monitoring that can detect potential misuse.
OpenAI's parallel struggle over GPT-5.6 created a precedent. On June 25, just a day before the Anthropic letter, OpenAI announced a similar limited access arrangement for its most powerful model. That agreement allowed select government agencies and vetted enterprises to use GPT-5.6 under strict monitoring conditions. The Anthropic deal mirrors this structure, with the added complication that Mythos 5's cybersecurity focus requires even tighter controls over who can deploy it.
Industry observers noted that the government's case-by-case approach has created uncertainty across the AI sector. Startups and established labs alike have complained that the lack of clear regulatory guidelines makes it impossible to plan long-term product roadmaps. Some have accused the administration of using export controls to favor politically connected firms.
Competitive Pressures and Global Implications
While Anthropic and OpenAI were restricted, other AI labs—both domestic and international—continued to advance. Chinese company Baidu's Ernie Bot 5.0 Cyber Defender achieved scores comparable to Mythos 5 on several key benchmarks, and the open-source community released a number of specialized cybersecurity models that could be deployed without government interference.
US intelligence agencies expressed frustration at losing access to Mythos 5 during a period of heightened cyber threats. Russian and Chinese state-sponsored hacking groups have intensified attacks on critical US infrastructure in recent months, and the NSA had relied heavily on the model to automate threat hunting.
Anthropic's board considered legal action against the government, but ultimately chose to negotiate rather than escalate. The company's investors, including Google and Salesforce, urged a cooperative approach to avoid damaging regulatory relationships.
What Happens Next?
For now, a small number of approved organizations—mostly federal agencies and key private sector partners like Microsoft and Amazon Web Services—will regain access to Mythos 5. Anthropic is working to restore service as quickly as possible, but the exact rollout timeline remains unclear. Fable 5, the model intended for general consumers and developers, is still blocked entirely. The government has given no indication of when or if it will be released.
The broader lesson from this saga is that AI regulation in the United States remains fragmented and unpredictable. While the Trump administration has taken a more hawkish stance than its predecessor, the lack of a comprehensive legal framework means ad-hoc decisions are made behind closed doors. Some industry groups are now pushing for a federal AI regulatory agency that would set clear rules for model release, similar to the FDA's role in drug approval.
Anthropic's experience also highlights the tension between safety and accessibility. The company's own stated mission—developing safe AI for the benefit of humanity—has been partially compromised by the inability to deploy its most powerful tools widely. CEO Amodei remarked in a recent interview: "We can't protect the world if we can't put our models in the hands of the people who need them most."
The partial restoration of Mythos 5 is a positive step, but it underscores how fragile the AI ecosystem remains. As competitors catch up and new threats emerge, the question of who gets to control cutting-edge AI—and under what conditions—will only become more urgent.
Source: The Verge News