Crypto Community Remains Cautious as Anthropic Launches Claude Mythos

Venture capitalist Simon Dedic said Anthropic’s newest AI models have reduced the expertise and expense required to uncover cryptocurrency vulnerabilities to what he described as “almost zero.”

AI firm Anthropic on Tuesday unveiled the first public release of its advanced Claude Mythos model, known as Fable 5, prompting concerns among some crypto users that the system could be misused for harmful activities despite the safeguards built into it.

Anthropic said last month that its Mythos model identified more than 10,000 high- and critical-severity security flaws across what it described as “systemically important software,” raising questions among observers about whether the technology should be made publicly available.

This came even as the company stated on Tuesday that Fable 5 had been prepared for broad public use and includes protective measures that redirect certain subjects, including cybersecurity-related queries, to a separate model known as Claude Opus 4.8.

“Releasing a model this capable comes with risks. Without safeguards, Fable 5’s capabilities in areas like cybersecurity could be misused to cause serious damage,” it said.

The safeguards have provided little comfort to many crypto users, as artificial intelligence is being increasingly leveraged in attacks targeting digital asset platforms. In April, cryptocurrency losses from hacks climbed to $629.7 million, the highest level recorded since February 2025, with analysts attributing part of the surge to the growing use of AI-driven techniques.

Mythos Launch Triggers Safety Concerns Among Crypto Users#

Simon Dedic, founder of venture capital firm Moonrock Capital, wrote on X on Tuesday that Fable 5 could reduce the expertise and resources needed to identify vulnerabilities in smart contracts to what he described as “virtually zero.”

“For DeFi, this should be a massive wake-up call. Unaudited protocols will become sitting ducks. Known exploits will get replayed on forks around the clock. Even small projects will get targeted simply because trying costs next to nothing now,” he added.

Dedic reiterated his warnings on social media, urging crypto holders to take protective measures such as revoking wallet permissions, withdrawing as much capital as possible from protocols, and transferring digital assets into newly configured hardware wallets.

However, Curve Finance co-founder Michael Egorov argued that concerns surrounding Claude Mythos may be exaggerated, noting that the model’s ability to uncover flaws in conventional software does not necessarily mean it will be equally effective at identifying smart contract weaknesses within DeFi protocols.

In May, Anthropic reported that Claude Mythos identified thousands of critical security flaws in key software systems through its Project Glasswing initiative. Among open-source projects, which play a major role in the operation of crypto protocols, the model uncovered roughly 6,200 high- or critical-severity vulnerabilities across more than 1,000 projects that it analyzed.

Egorov maintained that the software examined by Mythos contained millions of lines of code, whereas smart contracts typically consist of only a few thousand. He added that both human developers and conventional AI systems can fully process that code within context and are able to reason through it effectively.

“I suspect we might not be having a wave of DeFi code hacks, but we may see a lot of things in OpSec [operational security] getting hacked (looking like multisig keys compromises) and supply chain attacks on frontend dependencies, and those are way less dangerous in true DeFi,” he said.

Meanwhile, Anthropic announced that a limited group of cybersecurity and infrastructure firms would receive access to Claude Mythos 5. The model shares the same foundation as Fable 5, though certain safeguards have been relaxed in selected areas to expand its capabilities.