[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-agents-design-better-materials-for-about-1-a-run":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},2686,"ai-agents-design-better-materials-for-about-1-a-run","AI Agents Design Better Materials for About $1 a Run","A new framework uses LLM agents to autonomously discover high-performing metal-organic frameworks without training a separate model for each target.","LLM agents can now run closed-loop materials discovery at roughly $1 per campaign — and explain their reasoning along the way.\n\nResearchers introduced LLM4MOF, a framework in which two language-model agents collaborate to design metal-organic frameworks (MOFs) — porous crystals used in gas storage, carbon capture, and electronics. One agent proposes hypotheses about which metal nodes, organic linkers, pore shapes, and functional chemistry might work best. A second translates those hypotheses into constraints that select or construct candidate structures. The loop runs for ten autonomous iterations, testing each hypothesis through four diagnostic configurations that isolate whether geometry, chemistry, or the metal choice is doing the work. Across six tasks covering gas adsorption, separation, and electronic structure, the system concentrated its search on top-performing structures within 400 property evaluations — without ever seeing the full property landscape of the databases it drew from.\n\nWhat separates this from prior ML-for-materials work is the interpretability. Most machine-learning models for MOF design are black boxes: they find good structures but cannot say why. LLM4MOF produces human-readable design hypotheses at every step, which means a chemist can audit, redirect, or learn from the process rather than just accepting a ranked list. At roughly $1 per campaign, the cost barrier to running many such experiments drops to nearly nothing.\n\nThe result is a pointed challenge to the standard playbook of training a dedicated surrogate model for each design objective — a process that is expensive, slow, and opaque. Whether LLM agents can hold up as MOF complexity scales is still an open question, but the price-to-insight ratio here is hard to argue with.","[\"ai\",\"materials-science\",\"research\",\"llm-agents\"]","2026-06-30T04:00:00.000Z","2026-06-30T10:53:22.659Z","2026-06-30T10:53:25.530Z","published",null,[],"ai",[24,26,27,28],"materials-science","research","llm-agents",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2606.29459",0,{"sections":35},[36,40,45,50,55,60,65,70,75,80,85,89,94,99],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":44},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":49},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Dev Tools","dev-tools",59,"2026-07-07T04:00:00.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":86,"slug":87,"count":83,"latest_published_at":88},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":93},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]