[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-study-finds-most-multi-agent-ai-setups-lose-to-a-single-model":10,"sections":35},{"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":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},11340,"study-finds-most-multi-agent-ai-setups-lose-to-a-single-model","Study Finds Most Multi-Agent AI Setups Lose to a Single Model","A new benchmark shows that stacking multiple LLM agents together usually burns extra tokens without beating a well-run single-agent baseline.","More agents, same or worse answers, for a lot more money.\n\nResearchers built a shared evaluation framework called BenchAgent and tested six multi-agent workflow setups against a single-agent baseline called BenchAgent Core, all running GPT-4.1 across ten reasoning, coding, and tool-use benchmarks. Over three runs covering 3,302 task instances each, only one configuration, EvoAgent, beat Core overall, hitting 79.70% mean accuracy versus Core's 76.65%. The other five lagged behind by 1.71 to 8.76 percentage points. Two of the laggards illustrate the tradeoff starkly: LLM-Debate improved on HumanEval by 9.67 points but dropped 1.96 points overall while burning 1.39 times Core's tokens, and ChatEval gained 22.25 points on IFEval while trailing Core overall and using 4.34 times the tokens.\n\nThis matters because \"multi-agent\" has become a default pitch in AI tooling, often sold as a free upgrade over a single model. This benchmark is a rare apples-to-apples check on that claim, using fixed step limits and shared scoring across ten tasks rather than one cherry-picked demo. The pattern it finds, big token bills for narrow or negative gains, is exactly the failure mode teams don't see until the invoice arrives.\n\nIf your multi-agent system can't beat a single well-configured model on your actual workload, you're not buying intelligence. You're buying API calls.","[\"multi-agent-systems\",\"llm-benchmarks\",\"gpt-4\",\"ai-research\"]","2026-10-12T04:00:00.000Z","2026-10-12T09:03:38.053Z","2026-10-12T09:03:43.642Z","published",null,[],"ai",[26,27,28,29],"multi-agent-systems","llm-benchmarks","gpt-4","ai-research",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2606.05670",0,{"sections":36},[37,40,45,50,55,60,65,70,75,80,85,90,95,100],{"name":38,"slug":24,"count":39,"latest_published_at":18},"AI",6966,{"name":41,"slug":42,"count":43,"latest_published_at":44},"Security","security",952,"2026-10-12T04:54:12.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":49},"Policy","policy",496,"2026-10-11T20:06:43.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Deals","deals",489,"2026-10-12T13:30:00.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Hardware","hardware",237,"2026-10-12T13:24:45.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Science","science",196,"2026-10-11T15:34:40.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Consumer Tech","consumer-tech",192,"2026-10-12T14:07:13.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Startups","startups",121,"2026-10-12T13:22:34.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Software","software",115,"2026-10-11T07:34:45.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Dev Tools","dev-tools",111,"2026-10-12T11:22:24.000Z",{"name":86,"slug":87,"count":88,"latest_published_at":89},"General","general",70,"2026-10-11T14:15:00.000Z",{"name":91,"slug":92,"count":93,"latest_published_at":94},"Gaming","gaming",59,"2026-10-09T11:43:43.000Z",{"name":96,"slug":97,"count":98,"latest_published_at":99},"Reviews","reviews",35,"2026-10-12T14:00:00.000Z",{"name":101,"slug":102,"count":103,"latest_published_at":104},"How-To","how-to",8,"2026-10-05T09:00:00.000Z"]