{
  "id": 16290496,
  "title": "Moonshot curbs hacking ability in new Kimi model as fears grow over AI-powered cyberattacks",
  "url": "https://urgent.news/2026/10/24/moonshot-curbs-hacking-ability-in-new-kimi-model-as-fears-grow-over-16290496",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-24T06:00:39.000Z",
  "source": {
    "name": "Investing.com",
    "slug": "investing-com",
    "url": "https://www.investing.com/news/stock-market-news/moonshot-curbs-hacking-ability-in-new-kimi-model-as-fears-grow-over-aipowered-cyberattacks-4970639"
  },
  "original_language": "en",
  "account": "Beijing-based Moonshot AI has reduced the cyber-offensive capabilities of its newest model, the Kimi K3.1 Pro, as growing fears emerge that autonomous AI could enhance hacking. The company introduced the 3-trillion-parameter-class open-weight model on Saturday. Moonshot plans to release the model's weights by November 4, which allow developers to run and modify the model independently. To curb cyber capabilities, Moonshot removed cyber-related data from the training corpus and reinforcement-learning environments used to develop Kimi K3.1 Pro, while maintaining general coding skills. The company stated it intends not to release open weights that would advance cyber capability to the highest level. China has proposed export controls on AI-driven malware-generation and mutation technology, restricting AI-assisted vulnerability discovery tools. Moonshot believes open-weight models can enhance security by allowing more organizations to inspect and secure their own software, including legacy systems that receive limited attention. However, the open weights make central oversight more challenging once widely distributed. On an internal cyber benchmark, K3.1 Pro scored 81.0%, compared to 90.5% for a similar version without risk-reduction measures. This test involved identifying and exploiting software vulnerabilities. Moonshot claims the filtering of cyber-related data reduced the model's offensive cyber performance, though the results have not been independently verified. K3.1 Pro features a 1-million-token context window, enabling it to process extensive amounts of text simultaneously. Additionally, the company released a faster version, K3.1 Pro Highspeed, which runs at two to three times the speed of the standard model.",
  "summary": null,
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 3,
    "also_reported_by": [
      {
        "outlet": "Economic Times Tech",
        "title": "Moonshot curbs hacking ability in new Kimi model as fears grow over AI-powered cyberattacks",
        "url": "https://urgent.news/2026/10/24/moonshot-curbs-hacking-ability-in-new-kimi-model-as-fears-grow-over-16289289",
        "published": "2026-10-24T05:12:30.000Z"
      },
      {
        "outlet": "The Business Times - Companies & Markets",
        "title": "Moonshot curbs hacking ability in new Kimi model as fears grow over AI-powered cyberattacks",
        "url": "https://urgent.news/2026/10/24/moonshot-curbs-hacking-ability-in-new-kimi-model-as-fears-grow-over",
        "published": "2026-10-24T05:54:38.000Z"
      }
    ]
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}