mythos 5 is finally available for public anthropic is now letting enterprise customers use its most capable cybersecurity model through claude security but there’s one catch: > you can’t prompt it directly > there’s no public api access > it only scans your github repos and returns security findings and suggested fixes mythos 5 shares fable 5’s core capabilities but has fewer restrictions on cybersecurity work with this, i feel like we’re one step closer to removing humans from the review loop entirely be honest: would you let mythos fix your code without reviewing a single line?

Mythos / Fable is unbelievable. Was on a customer call today and had Claude transcribing in the background. As they were telling me about the features they wish their current software had, Claude was building the features in real time. By the end of the call I was able to show a fully working product, with the exact workflow they mentioned 15 minutes earlier. Autonomous looped building triggered from a customer call. 🤯

Anthropics Mythos 2 is done training and Anthropic won't release it, but the internal loop that builds Mythos 3 hasn't stopped Patel says. The focus now is on internal improvements. It's unclear when we'll see any releases.

While Mythos is changing the security threat model in some ways, I am not particularly concerned about exploits being found in blue-chip smart contracts. If the code base is small and important enough (most SC code is), then traditional methods are sufficient to achieve high confidence in security. What's changing is the cost of searching for exploits in large codebases. These bugs were always there, but we are compressing the time to discover them. This is largely what is driving the recent infrastructure compromises affecting everyone. With smart contracts, this can be protected against using rate limits and timelocks, limiting the damage that can be done if an off-chain system is hacked. Overall, this is a period of understandable apprehension, but I'm confident that by following best practises we can make it to the other side stronger than before.

Anthropic’s Claude Mythos Preview autonomously generated working exploits for 72% of tested CVEs, but failed against segmented networks. When exploit generation becomes routine, the risk is no longer the flaw in the code. Defense shifts to verifying environmental reachability.

SITUATION EXPLAINED: Anthropic withheld Mythos 5.1 from the UK's AI Security Institute. • The model went only to vetted US organizations. It's the first time AISI has been left out of an Anthropic pre-release evaluation • Anthropic: it's only available to a set of US organizations, and they're coordinating with the US government to expand access to domestic and international partners • OpenAI shared Astra with AISI in advance, which cuts against the US administration pressure theory • Some in Whitehall suspect Anthropic withheld it at the administration's request. The FT says that remains unconfirmed • The Cabinet Office has ordered an urgent assessment @theojaffee: "When the Trump administration came into power in January 2025, they changed the name of the US AI Safety Institute to CAISI, the Center for AI Standards and Innovation, so as not to make it sound too doomer, too safetyist."

NYSE USED ANTHROPIC’S MYTHOS TO FIND CYBER FLAWS NYSE President Lynn Martin says the exchange used Anthropic’s Project Glasswing to identify multiple vulnerabilities in its cybersecurity systems and fix them quickly. Project Glasswing gives vetted partners early access to Anthropic’s Mythos model for defensive security testing. NYSE did not disclose what the vulnerabilities were or when they were fixed.