Is Anthropic’s Next AI Model a Threat to Cybersecurity?

Anthropic is a company that is in business to develop the capabilities of AI. As I’ve described here and here, there has been some level of panic about the degree to which AI in general is moving faster than anyone expected. In particular, people have cited Anthropic’s tool known as Claude Code as essentially able to independently complete hours of coding work in a way that renders a lot of junior programmers obsolete.

“We’re talking 10 to 20 — to even 100 — times as productive as I’ve ever been in my career,” Steve Yegge, a veteran coder who built his own tool for running swarms of coding agents, told me. “It’s like we’ve been walking our whole lives,” he says, but now they have been given a ride, “and it’s fast as [expletive].”

In theory this is great and could lead to a boom in production, but there’s an obvious potential downside to having a tool like Claude Code that can do all of this better and faster than a lot of humans. In the wrong hands, this tool could absolutely wreck the cybersecurity of a lot of big institutions.

Now we get to what is either the genuinely scary part of this story or some genuinely brilliant PR work by Anthropic. The company says that it’s new model, which is called Mythos, is so good at coding that it can almost instantly find (and in theory exploit) vulnerabilities in code, some that have existed unnoticed for decades. Anthropic is so worried about the potential for this to go badly that it is essentially holding the new model back from public release.

One balmy February evening in Bali, Nicholas Carlini stepped away between events at a wedding, opened his laptop, and set out to do some damage. Anthropic PBC had just made a new artificial intelligence model, called Mythos, available for internal review, and Carlini — a well-known AI researcher — intended to see what kind of trouble it could cause.

Anthropic pays Carlini to stress-test its AI models to see whether hackers could leverage them for espionage, theft or sabotage. From Bali, where Carlini and his wife were attending an Indian wedding, he was staggered at what the model could do.

Within hours Carlini found numerous techniques to infiltrate systems used around the world. Once Carlini was back in Anthropic’s downtown San Francisco office, he discovered Mythos was able to autonomously create powerful break-in tools, including against Linux, the open-source code that underpins most of modern computing…

A previous model, Opus 4.6, had shown indications it could help people exploit vulnerabilities in software. Mythos could exploit the vulnerabilities on its own, Graham says. This was a national security risk, he warned Anthropic’s executives. That left Graham with the unenviable task of telling his bosses that their next major revenue generator was too hazardous to release to the public.

What the company eventually decided to do, was a kind of pre-release to insiders which would allow them to use Mythos itself to, essentially, Mythos-proof their own systems.

…rather than make it widely available to Claude users, Anthropic gave 12 tech companies access via Project Glasswing, which it described as “an effort to secure the world’s most critical software”.

They include cloud computing giant Amazon Web Services, device manufacturers Apple, Microsoft and Google, and chip-makers Nvidia and Broadcom.

Crowdstrike, whose faulty software update caused a major global outage in July 2024, is also among the project’s partners, with Anthropic saying it has also given access to Mythos to more than 40 organisations responsible for critical software.

In a video released alongside Project Glasswing’s launch, Anthropic boss Dario Amodei said it had offered to work with US government officials to “help defend against the risk of these models”.

Not everyone is convinced the Mythos threat is real, but David Sacks, the White House AI advisor says it has to be taken seriously because Anthropic’s claims could be real.

David Sacks: We have no choice but to take the Mythos threat seriously

“Anytime Anthropic is scaring people, you have to ask, is this a tactic, is this part of their chicken little routine, or is it real?

With cyber, I actually would give them credit in this case and say this… pic.twitter.com/2Wwjazc3Ck

— The All-In Podcast (@theallinpod) April 15, 2026

Even beyond our borders, some experts are worried.

Ciaran Martin, former head of the UK’s National Cyber Security Centre, told the BBC earlier this week the claim Mythos could unearth critical vulnerabilities much more quickly than other AI models had “really shaken people”.

“The second thing is that even with existing weaknesses that we know about, but organisations might not have patched against, might not be well defended against, it’s just a really good hacker,” he said.

You may also remember that Anthropic has been in a big fight with the Pentagon after it refused to allow the military to use Claude for AI guided weapons. The Pentagon responded by blacklisting the company and now the fight is taking place in court after Anthropic sued. 

However, the founder of Anthropic, Dario Amodei is heading the White House today for what Axios has dubbed “peace talks” so we may see something change soon.

Reminder: Anthropic is suing the Pentagon for blacklisting the company after Amodei refused to allow his AI to be used without restrictions.

Some parts of the U.S. intelligence community, plus the Cybersecurity and Infrastructure Security Agency (CISA, part of Homeland Security), are testing Mythos. Treasury and others want it.

Behind the scenes: After Anthropic took the administration to court, negotiations with the Pentagon chilled. But Anthropic has hired key Trumpworld consultants — so expect a deal. Friday’s meeting is designed to pave the way.

We’ll have to wait and see but, ultimately, Mythos may be too valuable to other parts of the government to allow the Pentagon to scuttle it for everyone.

Editor’s Note: Do you enjoy Hot Air’s conservative reporting that takes on the radical left and woke media? Support our work so that we can continue to bring you the truth.

Join Hot Air VIP and use promo code FIGHT to receive 60% off your membership.