How does MAI-Thinking-1 compare to other reasoning models?

Microsoft says MAI-Thinking-1, a medium-sized model, "matches leading models" on key software engineering benchmarks. The company trained it from the ground up on clean data without distillation from third-party models.

Which new Microsoft AI models were announced?

Microsoft announced MAI-Thinking-1 (reasoning), MAI-Image 2.5 and its flash version (text-to-image and image editing), MAI-Transcribe-1.5 (transcription, five times faster than competing models), MAI-Voice-2 with 15 new languages, and MAI-Code-1-Flash (coding, integrated into GitHub Copilot and Visual Studio Code).

When will MAI-Voice-2's flash version be available?

Microsoft says the flash version of MAI-Voice-2 is "coming soon" but did not announce a specific release date.

Back to articles

Microsoft launches MAI-Thinking-1, its new flagship reasoning AI model

The Verge AIJun 2, 2026

Key takeaway

Microsoft announced MAI-Thinking-1, a new flagship reasoning AI model developed in-house, alongside six other specialized models for image generation, transcription, voice, and coding at its Build 2026 conference. The move reflects Microsoft's strategic shift toward building proprietary AI models as it loosens its partnership with OpenAI, positioning the company to reduce dependence on third-party model providers.

Summaries like this, in your inbox every morning.

3 Key Points

What happened
Microsoft announced seven new in-house AI models at Build 2026, including MAI-Thinking-1, a "medium-sized" reasoning model the company says "matches leading models" on key software engineering benchmarks. The company also introduced MAI-Image 2.5, MAI-Transcribe-1.5 (described as "five times faster than competing models"), MAI-Voice-2 with 15 new languages, and MAI-Code-1-Flash for coding tasks.
Why it matters
This represents Microsoft's continued shift toward building its own AI models rather than relying solely on OpenAI—the two companies recently renegotiated their partnership to loosen ties. Microsoft trained MAI-Thinking-1 "from the ground up on clean data, without distillation from third-party models," signaling the company's investment in developing independent AI capabilities.
What to watch
MAI-Voice-2's flash version is "coming soon," and MAI-Code-1-Flash is already integrated into GitHub Copilot and Visual Studio Code, giving developers immediate access to the new coding model.

Get news like this every morning — free Read Original Article

FAQ

How does MAI-Thinking-1 compare to other reasoning models?: Microsoft says MAI-Thinking-1, a medium-sized model, "matches leading models" on key software engineering benchmarks. The company trained it from the ground up on clean data without distillation from third-party models.
Which new Microsoft AI models were announced?: Microsoft announced MAI-Thinking-1 (reasoning), MAI-Image 2.5 and its flash version (text-to-image and image editing), MAI-Transcribe-1.5 (transcription, five times faster than competing models), MAI-Voice-2 with 15 new languages, and MAI-Code-1-Flash (coding, integrated into GitHub Copilot and Visual Studio Code).
When will MAI-Voice-2's flash version be available?: Microsoft says the flash version of MAI-Voice-2 is "coming soon" but did not announce a specific release date.

Get AI news like this every morning

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Subscribe free →

Free · takes 30 seconds · unsubscribe anytime

Discussion

No discussion yet for this article

Stay ahead with AI news

Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.

Get Started Free

Free · takes 30 seconds · unsubscribe anytime

1 minute a day. The AI essentials.

200+ sources · Email / LINE / Slack

Get it free →