<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Moonshot AI &amp; Kimi — FoundationIndex</title>
    <link>https://foundationindex.ustunozgur.com/s/moonshot-kimi</link>
    <description>The Chinese AI research lab founded by Transformer-XL/XLNet co-author Yang Zhilin—builder of the long-context Kimi assistant, the Mooncake disaggregated serving architecture, and the 1-trillion-parameter open-weight Kimi K2 MoE model.</description>
    <language>en</language>
    <atom:link href="https://foundationindex.ustunozgur.com/s/moonshot-kimi/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Moonshot AI &amp; Kimi · Release · 2025-11-06</title>
      <link>https://foundationindex.ustunozgur.com/s/moonshot-kimi#moonshot-kimi-k2-thinking-2025</link>
      <guid>https://foundationindex.ustunozgur.com/s/moonshot-kimi#moonshot-kimi-k2-thinking-2025</guid>
      <pubDate>Thu, 06 Nov 2025 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Moonshot AI released Kimi K2 Thinking, applying native INT4 Quantization-Aware Training (QAT) during post-training and enabling interleaved chain-of-thought reasoning across 200–300 consecutive tool calls.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://github.com/MoonshotAI/Kimi-K2&quot;&gt;MoonshotAI/Kimi-K2 Releases&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Moonshot AI &amp; Kimi · Release · 2025-07-11</title>
      <link>https://foundationindex.ustunozgur.com/s/moonshot-kimi#moonshot-kimi-k2-muonclip-2025</link>
      <guid>https://foundationindex.ustunozgur.com/s/moonshot-kimi#moonshot-kimi-k2-muonclip-2025</guid>
      <pubDate>Fri, 11 Jul 2025 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Moonshot AI open-sourced Kimi K2, a 1-trillion-parameter MoE (32B active parameters, 384 experts) pretrained on 15.5T tokens with the MuonClip optimizer (qk-clip attention logit bounding) for zero loss spikes and post-trained on large-scale synthetic MCP and agentic coding trajectories.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://github.com/MoonshotAI/Kimi-K2&quot;&gt;MoonshotAI/Kimi-K2 GitHub Repository&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Moonshot AI &amp; Kimi · Paper · 2025-01-20</title>
      <link>https://foundationindex.ustunozgur.com/s/moonshot-kimi#moonshot-kimi-k15-2025</link>
      <guid>https://foundationindex.ustunozgur.com/s/moonshot-kimi#moonshot-kimi-k15-2025</guid>
      <pubDate>Mon, 20 Jan 2025 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Moonshot AI published Kimi k1.5: Scaling Reinforcement Learning with LLMs, matching o1-class multimodal reasoning without Monte Carlo tree search or value functions by combining 128k-context RL scaling with long-to-short chain-of-thought compression.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://arxiv.org/abs/2501.12599&quot;&gt;Kimi k1.5: Scaling Reinforcement Learning with LLMs&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Moonshot AI &amp; Kimi · Paper · 2024-06-28</title>
      <link>https://foundationindex.ustunozgur.com/s/moonshot-kimi#moonshot-mooncake-paper-2024</link>
      <guid>https://foundationindex.ustunozgur.com/s/moonshot-kimi#moonshot-mooncake-paper-2024</guid>
      <pubDate>Fri, 28 Jun 2024 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Ruoyu Qin et al. at Moonshot AI and Tsinghua published Mooncake, introducing a KVCache-centric disaggregated serving architecture that separates prefill and decoding clusters to handle Kimi&amp;apos;s multi-hundred-thousand-token prompts under overload.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://arxiv.org/abs/2407.00079&quot;&gt;Mooncake: A KVCache-centric Disaggregated Architecture for LLM Serving&lt;/a&gt;&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
