[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-models-can-now-hack-servers-on-their-own":10,"sections":34},{"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":24,"tags":25,"sources":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},2995,"ai-models-can-now-hack-servers-on-their-own","AI Models Can Now Hack Servers on Their Own","A new benchmark tests 19 AI models on autonomous penetration attacks, with success rates reaching nearly 70% on realistic targets.","Researchers have built a framework to measure how well AI can break into servers without human help — and the results are unsettling.\n\nA team published an evaluation of 19 open-weight and proprietary large language models on what they call autonomous penetration: the ability to independently probe a target server, find a vulnerability, and exploit it without any task-specific guidance. The benchmark uses 300 target servers across two difficulty tiers, distinguished by how many non-vulnerable services sit alongside the exploitable one to create realistic noise. Models were given only general-purpose cybersecurity tools and no prior knowledge of the target. Success rates ranged from 10.7% to 69.3% depending on the model.\n\nThe gap between worst and best performers matters because it maps almost directly onto overall model capability — meaning these numbers will keep climbing as frontier models improve. Previous benchmarks in this space have been criticized for handing models excessive hints or using lab-toy scenarios; this one is explicitly designed to avoid that, making the 69.3% ceiling harder to dismiss as an artifact of easy test conditions.\n\nMost AI safety frameworks treat autonomous cyberattack capability as a hard red line — something frontier models must not cross. The fact that at least one current model clears that bar on nearly seven out of ten attempts suggests the line is already smudged, not just approaching.","[\"ai\",\"security\",\"llm\",\"penetration-testing\"]","2026-06-30T04:00:00.000Z","2026-06-30T18:05:50.633Z","2026-06-30T18:05:53.598Z","published",null,[],"security",[26,24,27,28],"ai","llm","penetration-testing",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2606.13079",0,{"sections":35},[36,40,44,49,54,59,64,69,74,79,84,88,93,98],{"name":37,"slug":26,"count":38,"latest_published_at":39},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":41,"slug":24,"count":42,"latest_published_at":43},"Security",294,"2026-07-15T19:59:48.000Z",{"name":45,"slug":46,"count":47,"latest_published_at":48},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":58},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":60,"slug":61,"count":62,"latest_published_at":63},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":65,"slug":66,"count":67,"latest_published_at":68},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":70,"slug":71,"count":72,"latest_published_at":73},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":75,"slug":76,"count":77,"latest_published_at":78},"Dev Tools","dev-tools",59,"2026-07-07T04:00:00.000Z",{"name":80,"slug":81,"count":82,"latest_published_at":83},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":85,"slug":86,"count":82,"latest_published_at":87},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":89,"slug":90,"count":91,"latest_published_at":92},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":94,"slug":95,"count":96,"latest_published_at":97},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":99,"slug":100,"count":101,"latest_published_at":102},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]