Microsoft's MAI Cyber model reports 50% adoption and 20% token efficiency gain
Overview
Mustafa Suleyman, Microsoft's AI chief, says on X that the MAI Cyber model used in Microsoft's security harness doubled user adoption to 50% over the last 30 days.
He also says the model is 20% more token efficient without any loss in quality. Microsoft has not published supporting data, and the post does not define the adoption metric or how quality was measured.
A later report from Testing Catalog says Microsoft has made MAI Cyber 1.1 Flash available on MDASH, its vulnerability-scanning platform. Testing Catalog reports that in an internal MDASH evaluation suite, the version matches MAI Cyber 1 Flash's performance while costing 18% less per scan. The report repeats Suleyman's 50% adoption figure, and the internal evaluation results have not been independently published.
Written by AI from the articles below · updated Oct 27, 1:26 PM ET
Check the sources:
Developments
2 developments
- Oct 27, 1:11 PM ET · 1 articleMicrosoft makes MAI Cyber 1.1 Flash available on MDASH for vulnerability scanningTesting Catalog: Microsoft releases MAI Cyber 1.1 Flash on MDASH
- Oct 27, 1:10 PM ET · 1 articleMustafa Suleyman says MAI Cyber model doubled adoption to 50% and is 20% more token efficientMustafa Suleyman: Microsoft's MAI Cyber model doubles adoption to 50% in 30 days
Article timeline
The articles in this story. Times are ET.
🚨 AI News | TestingCatalog@testingcatalogXMicrosoft releases MAI Cyber 1.1 Flash on MDASHAIMicrosoft makes MAI Cyber 1.1 Flash available on MDASH, its vulnerability-scanning platform. In an internal MDASH evaluation suite, the model matches MAI Cyber 1 Flash's performance while costing 18% less per scan. Microsoft AI CEO Mustafa Suleyman says the MAI Cyber model's user adoption in its security harness doubled to 50% in the last 30 days.

Mustafa Suleyman@mustafasuleymanXMicrosoft's MAI Cyber model doubles adoption to 50% in 30 daysAIMustafa Suleyman says Microsoft's MAI Cyber model, used in its security harness, has doubled user adoption to 50% over the last 30 days. He adds that the model is 20% more token efficient without any loss in quality.

Heat Trend
- Current comparable heat
- 12
- Peak comparable heat
- 20Oct 27, 2:00 PM ET
- Change in past 24 hours
- –
The trend compares only continuously observed participants, so its scope may be smaller than the current heat total. Hover or tap for hourly heat; use arrow keys with a keyboard.