<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>MAI — FoundationIndex</title>
    <link>https://foundationindex.ustunozgur.com/s/mai</link>
    <description>Microsoft AI&apos;s first-party family of frontier models spanning reasoning (MAI-Thinking-1), agentic coding (MAI Code 2), cybersecurity (MAI Cyber 2 Flash), real-time voice (MAI Realtime 1 &amp; MAI-Voice-2), transcription (MAI-Transcribe-2), and image generation (MAI-Image-2.7).</description>
    <language>en</language>
    <atom:link href="https://foundationindex.ustunozgur.com/s/mai/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>MAI · Release · 2026-11-17</title>
      <link>https://foundationindex.ustunozgur.com/s/mai#mai-code-2-cyber-2-realtime-1-ignite-2026</link>
      <guid>https://foundationindex.ustunozgur.com/s/mai#mai-code-2-cyber-2-realtime-1-ignite-2026</guid>
      <pubDate>Tue, 17 Nov 2026 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;At Microsoft Ignite 2026, Microsoft AI launched MAI Code 2 and MAI Code 2 Flash for repository-scale agentic coding in GitHub Copilot and Microsoft Foundry, MAI Cyber 2 Flash (a 1M-token context window model for automated vulnerability triage and patching), MAI Realtime 1 (native speech-to-speech at $0.01/minute), MAI Image 2.7, and MAI Licensed Weights for private self-hosted deployments in customer Azure tenants.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://microsoft.ai/news/introducing-mai-code-2-and-mai-code-2-flash/&quot;&gt;Introducing MAI Code 2 and MAI Code 2 Flash&lt;/a&gt;, &lt;a href=&quot;https://microsoft.ai/news/introducing-mai-cyber-2-flash/&quot;&gt;Introducing MAI Cyber 2 Flash&lt;/a&gt;, &lt;a href=&quot;https://microsoft.ai/news/introducing-mai-realtime-1/&quot;&gt;Introducing MAI Realtime 1&lt;/a&gt;, &lt;a href=&quot;https://microsoft.ai/news/introducing-mai-licensed-weights/&quot;&gt;Introducing MAI Licensed Weights&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>MAI · Release · 2026-11-12</title>
      <link>https://foundationindex.ustunozgur.com/s/mai#mai-transcribe-2-image-26-2026</link>
      <guid>https://foundationindex.ustunozgur.com/s/mai#mai-transcribe-2-image-26-2026</guid>
      <pubDate>Thu, 12 Nov 2026 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Microsoft AI released MAI-Transcribe-2 for multilingual speech-to-text diarization and streaming transcription alongside MAI-Image-2.6, advancing the quality-cost Pareto frontier on visual and audio benchmarks ahead of Ignite 2026.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://microsoft.ai/news/mai-transcribe-2-is-the-fastest-most-accurate-and-cheapest-speech-recognition-model-in-the-world/&quot;&gt;MAI-Transcribe-2 is the fastest, most accurate, and cheapest speech recognition model in the world&lt;/a&gt;, &lt;a href=&quot;https://microsoft.ai/news/pushing-the-quality-cost-frontier-with-mai-image-2-6/&quot;&gt;Pushing the quality-cost frontier with MAI-Image-2.6&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>MAI · Release · 2026-06-02</title>
      <link>https://foundationindex.ustunozgur.com/s/mai#mai-thinking-1-build-2026</link>
      <guid>https://foundationindex.ustunozgur.com/s/mai#mai-thinking-1-build-2026</guid>
      <pubDate>Tue, 02 Jun 2026 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;At Microsoft Build 2026, Microsoft AI launched seven in-house models in Microsoft Foundry anchored by MAI-Thinking-1—a ~1-trillion-total / 35-billion-active parameter sparse MoE reasoning model trained from scratch without third-party distillation—alongside MAI-Code-1 Flash, MAI-Cyber-1 Flash, MAI-Image-2, MAI-Voice-2, and MAI-Transcribe-1.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://microsoft.ai/news/introducing-mai-thinking-1/&quot;&gt;Introducing MAI-Thinking-1&lt;/a&gt;, &lt;a href=&quot;https://microsoft.ai/news/building-a-hillclimbing-machine-launching-seven-new-mai-models/&quot;&gt;Building a Hill-Climbing Machine: Launching Seven New MAI Models&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>MAI · Release · 2025-10-13</title>
      <link>https://foundationindex.ustunozgur.com/s/mai#mai-image-1-2025</link>
      <guid>https://foundationindex.ustunozgur.com/s/mai#mai-image-1-2025</guid>
      <pubDate>Mon, 13 Oct 2025 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Microsoft AI unveiled MAI-Image-1, its first in-house text-to-image model designed for photorealistic lighting and fast creative iteration in Copilot and Bing Image Creator.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://microsoft.ai/news/introducing-mai-image-1-debuting-in-the-top-10-on-lmarena/&quot;&gt;Introducing MAI-Image-1, debuting in the top 10 on LMArena&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>MAI · Release · 2025-08-28</title>
      <link>https://foundationindex.ustunozgur.com/s/mai#mai-1-preview-voice-1-2025</link>
      <guid>https://foundationindex.ustunozgur.com/s/mai#mai-1-preview-voice-1-2025</guid>
      <pubDate>Thu, 28 Aug 2025 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Microsoft AI introduced its first two in-house foundation models: MAI-Voice-1 for expressive low-latency speech generation in Copilot Daily and Podcasts, and MAI-1-preview, a sparse Mixture-of-Experts language model trained across ~15,000 NVIDIA H100 GPUs.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://microsoft.ai/news/two-new-in-house-models/&quot;&gt;Two new in-house models in service of people&lt;/a&gt;&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
