Microsoft has introduced Project Perception, an AI-driven cybersecurity system combining MAI-Cyber-1-Flash and GPT-5.4 that outperforms Anthropic’s Claude Mythos 5 on industry-standard security benchmarks.

Key Points

• Project Perception enters public preview on August 3 and will be integrated directly into Microsoft Defender.
• The system achieved a 96% score on the CyberGym benchmark, surpassing the 84% score recorded by Anthropic’s Claude Mythos 5.
• MAI-Cyber-1-Flash processes 90% of security queries, while the larger GPT-5.4 model handles the remaining 10% of complex tasks.
• Pricing is based on consumption of security compute units, with Microsoft claiming a 50% cost reduction compared to the existing MDASH configuration.
• The initiative aims to leverage Microsoft’s extensive enterprise data and cybersecurity expertise to defend against AI-powered threats.

Why it Matters

This development signals a shift toward more efficient, multi-model AI architectures in enterprise security, offering higher performance at a lower operational cost. By integrating these tools into Microsoft Defender, the company is positioning its platform as a primary defense against increasingly sophisticated AI-driven cyberattacks.