< Back to all clusters
[TECHNOLOGY] · United States · 17 sources

Microsoft unveils MAI‑Cyber‑1‑Flash AI model and Project Perception security platform

Microsoft announced the launch of MAI‑Cyber‑1‑Flash, its first foundation model built specifically for cybersecurity. Integrated into the MDASH vulnerability‑identification harness, the model achieved roughly a 96 % score on the CyberGym benchmark, outperforming Anthropic’s Mythos, Google’s Gemini and OpenAI’s GPT models, while operating at about half the cost of leading AI security solutions.

At the same event Microsoft introduced Project Perception, an agentic platform that coordinates specialized red, blue and green AI agents to simulate attacks, detect threats, investigate incidents and automate remediation. The platform is designed to handle up to 90 % of routine security tasks, passing the remaining complex cases to larger models such as GPT‑5.4. Project Perception will enter public preview in early August and is initially limited to vetted Azure customers.

Executives highlighted the strategic shift: Mustafa Suleyman, co‑founder of DeepMind and head of Microsoft AI, presented the announcements, and Hayete Gallot, Microsoft’s vice‑president for security, said Project Perception helps “defend against AI‑enabled attacks” at the speed and scale of modern threat actors.

Entities: Hayete Gallot · MAI-Cyber-1-Flash · MAI‑Cyber‑1‑Flash · Microsoft · Microsoft Corporation · Mustafa Suleyman · Project Perception

Claims

What the coverage asserts, and how well corroborated each claim is across sources.

  • [● 3 SOURCES] Microsoft says MAI‑Cyber‑1‑Flash outperforms competitors on the CyberGym benchmark, scoring about 96 %. (Microsoft)
  • [● 2 SOURCES] Hayete Gallot, Microsoft’s vice president for security, said Project Perception helps defend against AI‑enabled attacks. (Microsoft)
  • [● 2 SOURCES] Mustafa Suleyman, co‑founder of DeepMind and head of Microsoft AI, announced the model. (Microsoft)
  • [● 12 SOURCES] MAI‑Cyber‑1‑Flash reduces operational costs by roughly 50 % compared with leading AI security models. (company claim)
  • [● 13 SOURCES] MAI‑Cyber‑1‑Flash is integrated with Microsoft’s MDASH harness for vulnerability identification and remediation. (company technical detail)
  • [● 12 SOURCES] Project Perception is an agentic security platform that coordinates multiple specialized AI agents (red, blue, green teams). (company description)
  • [● 16 SOURCES] Microsoft launched its first cybersecurity‑focused foundation model named MAI‑Cyber‑1‑Flash. (multiple)
  • [● 4 SOURCES] Project Perception can simulate cyberattacks, detect threats, investigate activity, triage incidents, recommend fixes and automate remediation. (Microsoft)
  • [● 6 SOURCES] The platform performs about 90 % of routine security tasks, delegating the remaining 10 % to a larger GPT‑5.4 model. (company description)
  • [● 9 SOURCES] MAI‑Cyber‑1‑Flash achieved a roughly 96 % score on the CyberGym benchmark. (benchmark result)
  • [● 8 SOURCES] Microsoft says the system outperforms Anthropic’s Mythos, Google’s Gemini and OpenAI’s GPT models on the CyberGym benchmark. (company claim)
  • [● 4 SOURCES] Microsoft announced the launch at a security event in San Francisco on Monday. (event announcement)

Sources

about 16 hours ago