MAI-Image-2.5 is here — now #3 on text-to-image and #2 on image-to-image Arena leaderboards, surpassing Nano Banana Pro. Leading image generation. Precise editing. Built for enterprise scale.
It delivers strong performance on H100s, enabling deployment on existing infrastructure.
MAI-Thinking-1 is our first in-house reasoning model developed from scratch that is competitive with models of similar size on STEM reasoning and coding tasks.
35B active/1T total MOE.
💻Coding: 52.8% on SWE Bench Pro competitive with Opus 4.6
🧐 Reasoning: 97% on AIME 25
🤝Preferred to Sonnet 4.6 on blind side-by-side tests
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 🧵
It’s time to move from renting intelligence to truly controlling your AI. Microsoft Frontier Tuning lets you take our models and make them uniquely your own, turning them from capable generalists to completely custom partners.
It starts with reinforcement learning environments (RLEs) that allow our models to learn directly from your workflows. Think of them as training gyms for AI. Here the agent learns your very specific processes, your standards, your way of working. It goes from off-the-shelf to hyper-adapted to exactly what you and your teams need. Those adaptations drive efficiency and performance, and your unique models can keep continually learning in your RLEs. This changes the nature of AI – and it changes the impact.
For example, within Microsoft we use our RLEs combined with our MAI models to climb towards the best agentic use cases for Excel. Our MAI tuned model is on par with GPT-5.4 on public and private benchmarks, while being up to 10X more efficient.
Only you control your agents made with Frontier Tuning. You keep the benefits of your hard-earned know-how, data and institutional knowledge. With us, the RLEs and the models you build in them become your moat.
This is distinct. It’s a new era. An era of AI that you control, on your terms. I think it’ll be a good one. More on the blog: https://t.co/v65eop5aHS
Energized to be back at Build with the developer community and share what we have been building. Today, we're introducing new @Windows capabilities designed to meet developers where they are with less friction and more flexibility, across local and cloud, any language, any framework. https://t.co/fT5KAjW8qr
MAI-Code-1-Flash is here! Built and optimized for GitHub Copilot. From quick fixes to complex engineering challenges, write better code with more return on token. Rolling out to GitHub Copilot individual users in Visual Studio Code in the model picker and under the default auto picker now.
We’re making it easier to bring collaborative agents to Microsoft Teams.
Developers can now build agents that participate directly in conversations, understand context across chats and meetings, and help teams move work forward.
The new Teams SDK and CLI simplify the path from idea to working prototype. Agents can use targeted messages, reactions, and threaded sessions to support real workflows.
More details here: https://t.co/Bqiw8LUCZA
A three‑month pilot in a Midwestern bottling plant shows what happens when AI moves beyond chat and into decision-making, where constraints shift, stakes are real, and answers must hold. https://t.co/sDGaEXepa8
In a world where every GPU matters, Managed Compute in Foundry Models gives you global reach with zero infrastructure overhead by routing your workloads to available capacity automatically and globally.
https://t.co/5hIToWmmIc
Today’s news all comes down to this: we’re putting our relentless hill-climbing machine at your service. From launching top tier models to helping you make them your own, our commitment as a platform company is to keep you at the absolute frontier. For all the details: https://t.co/v65eop5Ixq
With the new MAI models and Frontier Tuning capabilities we announced today, we're focused on helping every company move from just consuming a frontier model to fully participating at the frontier.
Build on Tap! Microsoft CTO @Kevin_Scott shares 5 ways we're confusing AI capability and AI reality, and how to bridge the gap: https://t.co/8IH94hCeBH #MSBuild
What if your apps and agents understood work like people do?
That’s the vision behind Work IQ.
The Work IQ API helps developers build agents that go beyond data—understanding context, intent, and organizational signals.
Preview now. GA starting June 16.
Read more: https://t.co/y4xEZyZ16w
Introducing Frontier Tuning—a new way to continuously improve agents using your organization’s data, processes, and real-world workflows.
Engage with Microsoft FDEs to get started.
Learn more: https://t.co/OGXHzLT8oA
Our goal for developers has always been simple: help them move faster, be more agile, stay in flow, and build with confidence. That's what we're announcing at #MSBuild this week.
We're making the experience of developing on @Windows better than ever with faster developer execution, a secure execution layer for agents, and unmetered intelligence that runs locally on the device.
This includes the new @surface RTX Spark Dev Box, which combines @nvidia RTX Spark with a developer-optimized Windows 11 experience designed to help developers build, test, and run AI workloads.
Stop stitching vendors together to run open models in production. Fireworks AI on Microsoft Foundry is GA. One endpoint. Enterprise SLAs. Zero setup.
Build your next production workload with Fireworks AI on Foundry: https://t.co/isqI03Ll13
An open, end-to-end trust stack for AI agents on any framework: run ASSERT to find where your agent fails policy, use ACS to place the right controls at the right checkpoints, then re-run ASSERT to confirm.
Evaluation → enforcement, closed loop: https://t.co/yn4RmclIr7