Microsoft announced on Monday a new, lower-cost AI model designed to detect and repair cybersecurity vulnerabilities [1].
The launch marks a strategic effort to rejuvenate Microsoft's cybersecurity business by providing a cost-effective defense against AI-driven threats [1, 5]. The initiative falls under the leadership of Hayete Gallot, a former Google executive who now heads the company's cybersecurity unit [1, 2].
Marketed as Project Perception, also referred to as MAI-Cyber-1-Flash, the system is integrated with OpenAI’s GPT-5.4 [1, 3]. The company said the model is designed to identify security gaps and implement repairs more efficiently than previous iterations of security software [3, 4].
Microsoft said the new model provides a significant performance advantage over its competitors [3]. According to company data, the system outscored Anthropic’s Claude Mythos 5 by 12 points on a key benchmark [3].
Cost efficiency is a primary pillar of the new offering. Microsoft said the system costs roughly 50% of competing solutions [3]. This pricing strategy aims to make advanced AI security tools accessible to a broader range of enterprises, while maintaining high performance standards [3, 4].
The rollout comes as the industry faces an increase in sophisticated attacks that utilize artificial intelligence to find exploits. By combining the reasoning capabilities of GPT-5.4 with a specialized cybersecurity framework, Microsoft intends to position itself as the primary defender for corporate digital infrastructure [1, 5].
“The system costs roughly half of competing solutions”
Microsoft is attempting to shift the cybersecurity landscape from high-cost, niche tools to scalable, AI-integrated services. By leveraging the GPT-5.4 integration and aggressive pricing, the company is not only targeting market share from rivals like Anthropic but is also attempting to stabilize its own security reputation under new leadership.
