[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-moonshot-cuts-hacking-data-from-kimi-k31-pro-open-release":10,"sections":41},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":30,"tags":31,"sources":36,"feedback":40,"feedback_at":22,"cost_usd":40,"total_tokens":40},12310,"moonshot-cuts-hacking-data-from-kimi-k31-pro-open-release","Moonshot Cuts Hacking Data From Kimi K3.1 Pro Open Release","Moonshot says Kimi K3.1 Pro nearly matches Claude Opus 5.5 at coding and had hacking-related data cut, but it has not published details.","Chinese AI lab Moonshot has released Kimi K3.1 Pro, and says it stripped hacking-related material from the model's training data to limit misuse.\n\nThe model is an upgrade to Kimi K3, which is open-weight: the trained model files can be downloaded, run and modified by anyone, rather than reached only through a company's servers. Moonshot said the full weights are due by November 4. The company also said K3.1 Pro nearly matches Claude Opus 5.5 on a coding test.\n\nBoth claims come from Moonshot, and neither comes with much detail. The company has not disclosed which coding test it used, the scores, how much data it removed, or how it defines \"hacking-related.\" Until it does, \"nearly matches\" is a marketing phrase, not a measurement.\n\nThe safety approach is the more interesting part. Most AI labs limit misuse with refusal training, which teaches a model to decline certain requests. That works reasonably well on a hosted service, where the lab controls the model. It is a weaker fit for open weights, because anyone with the files can fine-tune a model to change how it behaves. Filtering the training data works at a different stage: it targets what the model learned, not what it has been told to decline.\n\nThat approach has a cost question attached. Security knowledge overlaps heavily with ordinary programming, and defenders use models to review code and understand attacks. A filter broad enough to matter could also dull the coding skills Moonshot is promoting, and one narrow enough to protect them may leave gaps.\n\nThe November 4 weights release is when outside testers can start checking both sides. For anyone weighing an open model for real work, the useful numbers will be how it scores on coding benchmarks and how it handles security tasks, not what the launch post says about either.","[\"moonshot\",\"open-weight\",\"ai-safety\",\"kimi\"]","2026-10-23T18:19:34.000Z","2026-10-23T19:02:50.384Z","2026-10-23T19:02:52.853Z","published",null,[24],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"editor-r1","editor",1,"The draft was written from a truncated source and says so: 'The source does not name the test or give scores' and 'at least in what we have' are process hedges that must go, so state plainly that Moonshot has not disclosed the test, scores, data volume or definition of 'hacking-related'. Also cut or soften the unsupported assertion that data filtering is 'sturdier' than refusal layers, and make sure the closing line is reader-facing analysis.","resolved","ai",[32,33,34,35],"moonshot","open-weight","ai-safety","kimi",[37],{"name":38,"url":39},"The Next Web","https:\u002F\u002Fthenextweb.com\u002Fnews\u002Fmoonshot-kimi-k3-1-pro-open-weight-cyber",0,{"sections":42},[43,47,52,57,62,67,72,77,82,87,92,97,102,107],{"name":44,"slug":30,"count":45,"latest_published_at":46},"AI",7075,"2026-10-23T19:32:39.000Z",{"name":48,"slug":49,"count":50,"latest_published_at":51},"Security","security",1063,"2026-10-23T19:41:05.000Z",{"name":53,"slug":54,"count":55,"latest_published_at":56},"Policy","policy",581,"2026-10-23T16:58:39.000Z",{"name":58,"slug":59,"count":60,"latest_published_at":61},"Deals","deals",569,"2026-10-23T14:43:38.000Z",{"name":63,"slug":64,"count":65,"latest_published_at":66},"Hardware","hardware",294,"2026-10-23T20:00:00.000Z",{"name":68,"slug":69,"count":70,"latest_published_at":71},"Consumer Tech","consumer-tech",293,"2026-10-23T18:42:54.000Z",{"name":73,"slug":74,"count":75,"latest_published_at":76},"Science","science",217,"2026-10-23T16:13:16.000Z",{"name":78,"slug":79,"count":80,"latest_published_at":81},"Startups","startups",148,"2026-10-23T16:26:29.000Z",{"name":83,"slug":84,"count":85,"latest_published_at":86},"Software","software",143,"2026-10-23T19:28:10.000Z",{"name":88,"slug":89,"count":90,"latest_published_at":91},"Dev Tools","dev-tools",122,"2026-10-21T16:00:00.000Z",{"name":93,"slug":94,"count":95,"latest_published_at":96},"General","general",101,"2026-10-23T18:46:30.000Z",{"name":98,"slug":99,"count":100,"latest_published_at":101},"Gaming","gaming",88,"2026-10-22T22:25:33.000Z",{"name":103,"slug":104,"count":105,"latest_published_at":106},"Reviews","reviews",38,"2026-10-20T12:00:00.000Z",{"name":108,"slug":109,"count":110,"latest_published_at":111},"How-To","how-to",8,"2026-10-05T09:00:00.000Z"]