Microsoft AI

Building a new class of safer, more capable AI systems we call Humanist Superintelligence: AI that is always aligned, controllable, and in service of humanity.

[Microsoft AI @MicrosoftAI]

Seven new models launching at Build: let’s go!
Reasoning. Code. Image. Transcribe. Voice.

Built from scratch on a clean data lineage, designed for efficiency, working seamlessly as a family of models

Thread 🧵
#MSBuild

[Microsoft AI @MicrosoftAI]

@TongHeConv Voice opens up a lot of possibilities. What would be your first use case?

[Microsoft AI @MicrosoftAI]

Inside one of the fastest growing AI labs in the world, a small team has been hard at work.

Reasoning. Code. Image. Voice. Transcription.

7 new models. Grounded in science and built on clean, commercially safe data. Designed to work seamlessly together.

ICYMI, watch @mustafasuleyman keynote at #MSBuild 2026 https://t.co/uAnbbYODLK

[Microsoft AI @MicrosoftAI]

@openingai_com It’s essentially training directly on your definition of success. What would you use it on first?

[Microsoft AI @MicrosoftAI]

Let’s go a bit deeper into Frontier Tuning launched at Build and see a live demo!

Frontier Tuning is how we enable you to develop custom AI by building a reinforcement learning environment (RLE) to hill climb on your data, tools, and knowledge. Think of it like a training gym for your AI.

[Microsoft AI @MicrosoftAI]

What does it take to build coding models that meet developers where they work? Go behind the scenes with Microsoft AI to explore how we build and optimize code. From training and evaluation to performance, safety, and real-world developer feedback. https://t.co/kkzNkci2ru

[Microsoft AI @MicrosoftAI]

MAI-Code-1-Flash is now rolled out to 100% of GitHub Copilot Free, Education, Pro, Pro+, and Max subscribers in VS Code. Copilot CLI roll out and Enterprise/Business preview on its way. Give it a try and let us know what you think!

[Microsoft AI @MicrosoftAI]

@AiLikeFive Good interpretation. It’s learning from your environment, not just data. What would you start with?

[Microsoft AI @MicrosoftAI]

@sahilyaps Glad that landed. Which model would you put to work first?

[Microsoft AI @MicrosoftAI]

@Microsoft @Azure @Copilot Learn more https://t.co/87B9KcOUKn

[Microsoft AI @MicrosoftAI]

See Frontier Tuning demo’d live at Build for MAI models across the @Microsoft stack, starting with @Azure and ending with @Copilot https://t.co/UNezm0eHav

[Microsoft AI @MicrosoftAI]

@AIVISIONOG Good observation. What would you create with it first?

[Microsoft AI @MicrosoftAI]

Fun fact – the images announcing MAI-Image-2.5 - our image generation and editing model were generated by the model itself!

Strong photo-realism, precise edits, quality down to the fine details.

Try it on our Playground: https://t.co/TipIcpmevb
or on @OpenRouter and share your favorite images on this thread

[Microsoft AI @MicrosoftAI]

@IamPranavJ We thought it was important to share the process, not just the results. Have you had a chance to read the paper?

[Microsoft AI @MicrosoftAI]

Seven new models launching at Build: let’s go!
Reasoning. Code. Image. Transcribe. Voice.

Built from scratch on a clean data lineage, designed for efficiency, working seamlessly as a family of models

Thread 🧵

[Microsoft AI @MicrosoftAI]

@aniruddhadak Interested to hear your take once you've tried it. What are you planning to create?

[James Montemagno @JamesMontemagno]

I have been crushing through some code including advanced research, planning, and implementation with MAI-Code-1-Flash and I am loving it.

Token usage is efficient and cost is fantastic. Flipping bugs for around 20 to 30 AIC each.

[Microsoft AI @MicrosoftAI]

@seymurglv Good use case to push it on. Let us know how it does if you try it.

[Microsoft AI @MicrosoftAI]

The best way to describe MAI-Voice-2 is to let it speak for itself.

15 languages with expressive range.

Listen 🎧👇

[Microsoft AI @MicrosoftAI]

@AiLikeFive That’s a good way to put it. Which part of the “toolbox” would you actually use first?

[Microsoft AI @MicrosoftAI]

Seven new models launching at Build: let’s go!
Reasoning. Code. Image. Transcribe. Voice.

Built from scratch on a clean data lineage, designed for efficiency, working seamlessly as a family of models

Thread 🧵
#MSBuild

[Microsoft AI @MicrosoftAI]

@BrandonWeger It’s a lot to explore. Which model are you most interested in?

[Microsoft AI @MicrosoftAI]

@CycleDecoded Sounds like you already have a test case in mind. What is it?

[Microsoft AI @MicrosoftAI]

Meet MAI-Transcribe-1.5

Our new speech-to-text model MAI-Transcribe-1.5 sits at the front of the Artificial Analysis Accuracy × Speed Pareto frontier. Up to 5x faster than other competitive models.

A thread on what's new 🧵