
Microsoft expands its AI model lineup across coding, security, image and voice tasks
The AMW Read
The release expands an established provider's model portfolio, but its claimed cost and performance advantages remain unverified.
Microsoft expands its AI model lineup across coding, security, image and voice tasks
Microsoft announced seven proprietary models at Ignite 2026, including MAI Code 2 and its lightweight Flash variant for GitHub Copilot. It also introduced MAI Cyber 2 Flash, two image models, a real-time voice model and Decision 1.1, which selects models according to task difficulty. Microsoft also announced Licensed Weights, which supports customization with enterprise data in dedicated cloud environments.
The release broadens Microsoft's role as a model supplier while it continues to distribute OpenAI and other providers' models. Task-specific models could give Microsoft more control over the cost of services delivered through its existing products. Microsoft says MAI Code 2 achieved 98% of GPT-6.1 Sol's average performance across four coding evaluations at about 40% of its cost. Those figures come from Microsoft's own tests and need independent verification.
For builders and investors, the key question is whether the claimed savings persist on their own workloads. Microsoft says MAI Cyber 2 Flash completed 92.3% of tasks in its security evaluation, versus 91.5% for OpenAI's GPT-6 Astra, at less than 2% of the cost per task. The comparison uses OpenAI's general-purpose model, not its security-specific model. Microsoft plans to offer MAI Cyber 2 Flash through its security products to selected eligible customers from mid-December.



