Home / Technology / Microsoft AI Outsmarts Rivals, Cuts Costs Dramatically
Microsoft AI Outsmarts Rivals, Cuts Costs Dramatically
28 Jul
Summary
- Microsoft's AI cybersecurity model MAI-Cyber-1-Flash achieves 96% on CyberGym benchmark.
- Combined with MDASH and GPT-5.4, it reportedly costs 50% less than current solutions.
- MAI-Cyber-1-Flash handles 90% of queries, delegating complex issues to GPT-5.4.
Microsoft has unveiled Project Perception, an advanced AI cybersecurity system that integrates its MAI-Cyber-1-Flash model with agentic system MDASH and OpenAI's GPT-5.4. This new offering aims to "use AI to defend against AI" and promises superior performance and cost efficiency.
In recent benchmarks on CyberGym, Microsoft's combined system achieved a 96% score. This notably surpasses the 84% score reported for Anthropic's Claude Mythos 5. The Project Perception system, entering public preview on August 3rd within Microsoft Defender, is set for a gradual rollout across all Microsoft Security products.
Mustafa Suleyman, CEO of Microsoft AI, explained that MAI-Cyber-1-Flash handles approximately 90% of queries, detecting and patching vulnerabilities. The remaining 10% of complex queries are escalated to the larger GPT-5.4 model. This collaborative approach reportedly delivers enhanced performance at approximately half the cost of current market configurations.
This development follows Anthropic's recent release of Claude Fable 5, which had faced scrutiny and necessary adjustments due to concerns about potential misuse. Microsoft attributes its AI advancements to its extensive data resources and the expertise of its cybersecurity professionals, positioning itself as a trusted steward of sensitive data.
