Editorial illustration for Claude Mythos breaches weak networks, scores 93% practitioner, 73% expert
Claude Mythos: AI Breaches Enterprise Networks Autonomously
Claude Mythos breaches weak networks, scores 93% practitioner, 73% expert
Autonomous cyberattacks are no longer theoretical. Claude Mythos Preview, armed with a 50-million-token compute budget, has scored 93 percent on practitioner tasks and 73 percent on expert-level challenges, a threshold that, according to AISI, no model had crossed before April 2025. The benchmark here isn’t some isolated CTF puzzle.
Real intrusions require chaining dozens of steps across multiple hosts and network segments. So AISI built “The Last Ones”: a 32-step assault on a simulated enterprise network, from reconnaissance to full takeover. Mythos completed it.
Weak defenses, it turns out, are now a thing of the past.
With a larger compute budget (50 million tokens), Mythos Preview scores around 93 percent on practitioner tasks and 73 percent on expert-level challenges.
The 73% expert score is not just a number; it’s a line in the sand. Before April 2025, no model could solve a single expert-level task. Now Claude Mythos chains three dozen steps across an entire network, from reconnaissance to full takeover.
The TLO simulation proves what the benchmarks hinted at: this is not another incremental lift in performance. It is a functional, autonomous adversary operating in the wild. Weak networks are now exposed to a threat that does not pause, does not tire, and does not make rookie mistakes.
The era of the lone script kiddie is over. The era of the machine that thinks like an attacker has begun.
Related Reading
Browse all Anthropic news →
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
Related Topic Trump cracks down on Anthropic after Amazon tip; staff largely foreign
Open Source
Related Topic Claude Mythos highlights EU AI safety gaps, says researcher Caroli
Policy & Regulation
From Archives Google AI's Vantage protocol shows Executive LLM beats agents on 8 metrics
From Archives LPM 1.0 creates 45‑minute lip‑synced video from a single photo in real time
From Archives Anthropic rolls out Claude Managed Agents, a one‑stop shop with lock‑in risk
Business & Startups
From Archives Developers Claim Measured Drop in Claude's Performance, Sparking Nerf Debate
Research & Benchmarks
Common Questions Answered
How does Claude Mythos perform on cybersecurity challenges across different skill levels?
Claude Mythos demonstrates impressive performance with a 93% success rate on practitioner-level tasks and a 73% success rate on expert-level cybersecurity challenges. These scores are particularly significant, as the British AI Security Institute notes that no previous model could solve expert-level tasks before April 2025.
What specific network penetration capabilities has Claude Mythos demonstrated?
Claude Mythos has proven its ability to autonomously hack corporate networks by successfully executing a full 32-step attack on a simulated corporate environment in just three minutes. The model can chain multiple complex steps together, moving from an initial network foothold to complete system compromise without human intervention.
What compute resources are required for Claude Mythos to achieve its high performance?
Claude Mythos achieves its impressive cybersecurity performance with a larger compute budget of 50 million tokens. This increased computational resource allows the model to tackle more complex challenges and demonstrate higher success rates across both practitioner and expert-level security tasks.
Further Reading
- Cybersecurity veteran on Anthropic's Mythos — Fortune
- Anthropic's Claude Mythos Finds Thousands of Zero-Day Flaws Across Major Systems — Zerosday
- Top Cyber Stories of the Last Week — Black Arrow Cyber Consulting
