Microsoft AI has released MAI-Cyber-1-Flash, a cybersecurity-specialized model that runs inside the MDASH agentic scanning harness to improve vulnerability detection. The model is a sparse Mixture-of-Experts transformer with 137B total and 5B active parameters, derived from the MAI-Code-1-Flash lineage.
- MAI-Cyber-1-Flash handles up to 90% of MDASH tasks, escalating the hardest 10% to GPT-5.4 for a claimed 50% cost saving.
- The combined system achieves 95.95% on CyberGym, an improvement from 88.45% when using generally available models in May 2026.
- The model scores zero on ExploitGym by design, as it is trained for defensive patching rather than offensive exploit generation.
- MDASH manages over 100 specialized agents across five stages to flag findings and prove exploitability.
The release allows Microsoft's Autonomous Code Security team to scale defense operations while significantly reducing the cost of using frontier models for routine scanning tasks.