What is Anthropic’s Fable 5 AI model and why is it causing concern in crypto?

Anthropic has released Fable 5, the first public version of its Claude Mythos AI model, which previously identified thousands of serious software vulnerabilities. Some crypto investors worry the model could make it easier and cheaper to discover weaknesses in DeFi protocols and smart contracts.

Why are crypto users worried about Anthropic’s Claude Mythos technology?

The concern is that advanced AI could automate parts of the vulnerability-hunting process that previously required specialised security expertise. Critics argue this could increase the number of attacks against smaller or poorly audited crypto projects.

How many vulnerabilities did Anthropic’s Mythos model find before its public release?

Anthropic said Mythos identified more than 10,000 high- and critical-severity vulnerabilities in systemically important software. The company also reported finding around 6,200 high- or critical-severity vulnerabilities across more than 1,000 open-source projects.

What did Moonrock Capital founder Simon Dedic say about Fable 5 and DeFi?

Simon Dedic warned that Fable 5 could reduce the cost and expertise needed to find exploitable smart contract flaws to “basically zero.” He argued that unaudited DeFi protocols could become easier targets if attackers use AI-powered security research tools.

Is Anthropic’s Fable 5 definitely a major threat to DeFi protocols?

Not necessarily. This remains a debated issue, with some industry participants viewing the risk as significant while others believe the threat is being overstated because smart contracts are generally smaller and easier to analyse than large software systems.

Why does Curve Finance co-founder Michael Egorov think the risks may be overblown?

Michael Egorov argues that smart contracts usually contain only a few thousand lines of code and are already reviewed by both human auditors and existing AI tools. He believes operational security failures and supply-chain attacks may be more likely attack vectors than direct smart contract exploits.

Image for illustrative purposes only. Not a real photo.

Anthropic AI release fuels crypto exploit fears

Written by Jon CuthbertPublished Jun 10, 2026, 5:12 AM

Anthropic has released Fable 5, the first public version of its Claude Mythos artificial intelligence model, prompting concerns among crypto users that the technology could be weaponised to identify vulnerabilities in decentralised finance protocols.

The launch follows Anthropic’s disclosure last month that Mythos uncovered more than 10,000 high and critical-severity vulnerabilities in systemically important software, raising questions about whether the model should be made publicly available.

“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,”

Said Anthropic.

Moonrock Capital founder Simon Dedic warned that Fable 5 could dramatically lower the barriers to discovering exploitable flaws in smart contracts, arguing that unaudited DeFi protocols could become increasingly vulnerable to attacks.

“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,”

Said Moonrock Capital founder, Simon Dedic.

Dedic also urged crypto users to revoke wallet approvals, reduce exposure to DeFi protocols and transfer assets to fresh hardware wallets as a precaution against potential threats enabled by advanced AI systems.

Curve Finance co-founder Michael Egorov disputed the severity of the threat, arguing that smart contracts typically contain only a few thousand lines of code and are already well understood by both human auditors and existing AI tools, while suggesting operational security failures and supply-chain attacks may present a greater risk than direct smart contract exploits.

Anthropic said a limited group of cybersecurity and infrastructure providers will receive access to Claude Mythos 5, a version of the model with some safeguards removed, while the company maintains that Fable 5 has been made safe for general use through built-in restrictions on sensitive cybersecurity topics.