Microsoft's MDASH Scores 95.95% on CyberGym, Beats Mythos 5 and GPT-5.6 Sol on July 27

MSFT1.93%
According to Microsoft, the company released MAI-Cyber-1-Flash, its first dedicated cybersecurity model, on July 27, achieving a 95.95% score on the CyberGym benchmark. The model paired with MDASH, Microsoft's vulnerability-hunting system, outperformed Anthropic's Mythos 5 (83.8%) and OpenAI's GPT-5.6 Sol (83.6%) while costing 50% less than Microsoft's previous best MDASH configuration. CyberGym evaluates AI agents' ability to reproduce known vulnerabilities across open-source projects. MAI-Cyber-1-Flash handles up to 90% of scanning workload, routing the most complex cases to GPT-5.4. Microsoft CEO Satya Nadella stated the combined setup "delivers world-class performance at half the cost." MDASH entered private preview through Microsoft Security Exposure Management, allowing customers to scan repositories and generate proposed code fixes.
Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments