Microsoft has announced MAI Cyber 1.1 Flash for verified MDASH customers. Its October 27 article reports comparable vulnerability-discovery performance to MAI Cyber 1 Flash with 18% lower cost per scan on an internal MDASH evaluation suite.
The system is part of the result
The announcement describes a model working within the MDASH harness, supported by Microsoft’s security data and workflow. The claimed improvement therefore concerns the evaluated scanning system, rather than a standalone token price that can be transferred directly to another product.
Microsoft says the more focused model performs more of the work itself. The stated goal is to reduce the cost of surfacing useful findings while retaining the performance defenders need. The article does not establish that every codebase will produce the same saving or the same findings.
Usage figures have a specific denominator
Microsoft reports that, after a customer’s first MAI-profile scan, nearly nine in ten subsequent scans use that profile rather than the OpenAI-only profile. It also says MAI Cyber accounts for almost half of MDASH scans.
Those figures describe scanning choices within the service. They do not mean nine in ten organisations use the product, establish market share or independently demonstrate the accuracy of a finding. Keeping the denominator attached makes the adoption claim easier to interpret.
Access through an existing workflow
The model is available to verified customers through MDASH. Existing customers can select its profile for a new scan; new users are directed to request access. Microsoft describes usage policies and ongoing monitoring as part of that access.
A defensive evaluation should compare accepted findings, false positives and review effort on authorised targets, alongside scan cost. The useful outcome is a verified issue that a team can address. A lower-cost scan is valuable when that review process still produces reliable evidence.