Microsoft CorporationMicrosoft launches its first cybersecurity AI model, MAI-Cyber-1-Flash, with strong benchmark performance.
Microsoft introduced MAI-Cyber-1-Flash, its first cybersecurity-specific AI model, built into its MDASH multi-agent vulnerability-hunting system. Paired with GPT-5.4, the combination scored 96% on the CyberGym benchmark, topping rivals from Anthropic, Google, and OpenAI by a wide margin. The new model handles about 90% of everyday security tasks, reserving GPT-5.4 for the toughest 10%, a routing approach Microsoft says cuts costs by 50% versus its previous MDASH offering. The launch comes alongside Project Perception, a new agentic security platform, and leverages over 100 trillion daily security signals from 1.6 million customers. Alphabet's DeepMind released its own cybersecurity model, Gemini 3.5 Flash Cyber, inside its CodeMender agent the same week, setting up a direct competition in AI-driven vulnerability defense.
Microsoft CorporationMicrosoft launches its first cybersecurity AI model, MAI-Cyber-1-Flash, with strong benchmark performance.
Alphabet Inc Class CMicrosoft's new cybersecurity AI model outperforms Alphabet's DeepMind model, intensifying competition.
DeepMind's cybersecurity model is outperformed by Microsoft's, indicating competitive pressure.
Microsoft's model tops rivals including Anthropic, suggesting competitive disadvantage.
OpenAI's GPT-5.4 is paired with Microsoft's model, but no direct impact on OpenAI is stated.