<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Ai2 (Allen Institute for AI) — USASI news</title>
    <link>https://unitedstatesofamericasuperintelligence.com/companies/ai2/</link>
    <atom:link href="https://unitedstatesofamericasuperintelligence.com/companies/ai2/feed.xml" rel="self" type="application/rss+xml"/>
    <description>Dated, sourced news items in the USASI catalog that mention Ai2 (Allen Institute for AI). Independent project. Not a United States government website.</description>
    <language>en-us</language>
    <lastBuildDate>Sun, 04 Oct 2026 12:00:00 GMT</lastBuildDate>
    <item>
      <title>Ai2 open-sources AstaBrief 8B, a model for writing cited scientific reports</title>
      <link>https://unitedstatesofamericasuperintelligence.com/news/ai2-open-sources-astabrief-8b/</link>
      <guid isPermaLink="true">https://unitedstatesofamericasuperintelligence.com/news/ai2-open-sources-astabrief-8b/</guid>
      <pubDate>Sun, 04 Oct 2026 12:00:00 GMT</pubDate>
      <category>release</category>
      <description>On October 2, 2026 Ai2 announced that it is open-sourcing AstaBrief 8B, a model that turns a research question and retrieved literature excerpts into a cited report, along with its training data. Ai2 says the model was built from Qwen3-8B using supervised fine-tuning and direct preference optimization, and that it is available as Fast mode in the &quot;Generate a report&quot; feature of Asta, its platform for scientific work. The Hugging Face model card lists the Apache 2.0 license. Ai2 notes that most of the training and evaluation was completed in 2025 and that it has not rerun the full evaluation against today's frontier models.</description>
    </item>
    <item>
      <title>Ai2 releases OLMo-core 3, its open training library, with mixture-of-experts support</title>
      <link>https://unitedstatesofamericasuperintelligence.com/news/ai2-releases-olmo-core-3/</link>
      <guid isPermaLink="true">https://unitedstatesofamericasuperintelligence.com/news/ai2-releases-olmo-core-3/</guid>
      <pubDate>Thu, 01 Oct 2026 12:00:00 GMT</pubDate>
      <category>release</category>
      <description>Ai2 released version 3 of OLMo-core, the PyTorch training library behind its Olmo models, on October 1, 2026. The release redesigns the library's mixture-of-experts training system, adding expert and pipeline parallelism, a distributed optimizer, grouped matrix multiplication for experts, and support for the MXFP8 number format; Ai2 says it is designed to scale mixture-of-experts training into the trillion-parameter range. The code is licensed under Apache 2.0, and the repository includes Ai2's official training scripts for Olmo 3.</description>
    </item>
  </channel>
</rss>
