# Microsoft launches seven MAI models at Ignite 2026

> Source: <https://www.testingcatalog.com/microsoft-launches-seven-mai-models-at-ignite-2026/>
> Published: 2026-11-17 19:16:22+00:00

Microsoft has introduced a family of seven new MAI models at Ignite 2026, aiming its first-party AI lineup at enterprises that want frontier-level results without frontier-level token bills. The release spans coding, security, image generation, voice, transcription and model routing, and is pitched at customers running agents around the clock.

MAI Code 2 and MAI Code 2 Flash lead the coding side. Microsoft says Code 2 reaches 98% of GPT-6.1 Sol's quality at about 40% of the price and 96% of Claude Opus 5.5 at 5% of the price, averaged across DeepSWE 1.1, Terminal Bench 2.1, SWE-Atlas and FrontierCode v1.1. It is tuned for everyday enterprise work and frontend output, though Microsoft notes it is not yet tuned for science research or very long-horizon agentic tasks. Code 2 Flash targets routine edits and fixes, matching GPT-6 Luna at 65% of the price and Claude Haiku 5.5 at 20%. Both are live in GitHub Copilot, with Flash also in limited preview on Microsoft Foundry.

MAI Cyber 2 Flash scores 92.3% on CyberGym-E2E against 91.5% for GPT-6 Astra at under 2% of the cost per task, and brings a one-million-token context window. It is priced at $0.10 input and $0.50 output per million tokens and is coming exclusively to Microsoft security products, replacing the current model in Codename MDASH for eligible customers from mid-December.

MAI Image 2.7 now ranks first on Artificial Analysis image editing, ahead of GPT Image 2.6 Sunburst at one fifth of the price, while MAI Image 2.7 Flash costs about $10 per 1,000 images against about $35 for the full model. Both are in public preview on Foundry and through partners including fal, Genspark, OpenRouter and Vercel, and Flash powers a new Canvas workspace in MAI Labs. Arena now lists Microsoft AI as the second-ranked lab across its three image leaderboards.

The rest of the lineup covers voice and routing. MAI Realtime 1 is Microsoft's first live speech-to-speech model, starting at $0.01 per audio minute, with interest registration open through MAI Labs. Dragon Copilot Transcribe, built on MAI Transcribe 2, already handles 10 million patient encounters a month. Microsoft-Decision-1.1, a routing model that sends requests to larger models only when needed, is generally available, and Artificial Analysis scores it 68 on its Decision Index, up from 62 for the previous version.

Microsoft is also launching MAI Licensed Weights, which lets enterprises customize a frontier MAI model with their own data and run it inside their own Azure environment, and says its Maia 200 accelerator is now in production running its largest models. Together, the launches position Microsoft AI as a lower-cost model supplier across Copilot, Foundry and Azure, competing with OpenAI and Anthropic mainly on price-performance.

## Sources and related context

- [MAI Code 2 and MAI Code 2 Flash in GitHub Copilot (GitHub changelog)](https://github.blog/changelog/2026-11-17-mai-code-2-and-mai-code-2-flash-in-github-copilot/?ref=testingcatalog.com) : Supports: GitHub confirms Code 2 for Copilot Pro, Pro+, Max, Business and Enterprise, and Code 2 Flash also for Free and Student plans, across VS Code, Visual Studio, CLI, JetBrains, Xcode and Eclipse.
- [Microsoft Ignite 2026 Book of News recap on Microsoft Source](https://news.microsoft.com/source/2026/11/17/microsoft-ignite-2026-empowering-human-ambition-intelligence-and-trust/?ref=testingcatalog.com) : Supports: Microsoft's corporate Ignite post confirms new coding, cyber, image and audio models and that Microsoft-Decision-1.1 is generally available.
- [Artificial Analysis on MAI-Image-2.7](https://x.com/ArtificialAnlys/status/2122754535206846610?ref=testingcatalog.com) : Supports: Artificial Analysis independently ranks MAI-Image-2.7 first on AA-Image-Editing v2.1, priced at $34.82 per 1,000 images, with Flash at $10.24.
