[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-still-cant-read-a-phase-diagram":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},3784,"ai-still-cant-read-a-phase-diagram","AI Still Can't Read a Phase Diagram","A new benchmark tests vision-language models on materials science phase diagrams — and finds they mostly fail at the hard parts.","Vision-language models can describe what a chart looks like. Understanding what it means is a different problem.\n\nResearchers built MatPhaseBench to find out how well current AI models handle materials phase diagrams — the dense, logic-heavy charts that encode temperature, composition, and phase stability for materials systems. The benchmark draws from 3,681 papers in classical materials science journals, narrowed to 200 high-quality diagram-text pairs covering 189 material systems and 70 elements. Every description was manually validated, not auto-generated. The tasks go beyond \"name what you see\" into open-ended questions that require thermodynamic reasoning.\n\nThe results are a reality check for anyone pitching AI as a near-term replacement for domain experts. Current models stay at the surface — identifying visual features — but fall apart when asked to reason about phase transformations or distinguish subtle differences between multi-panel diagrams. That gap between perception and mechanistic understanding is exactly where materials discovery work happens.\n\nBenchmarks like this matter because they expose the difference between \"impressive on general science questions\" and \"useful in a specialized research lab.\" Materials science is not unique here — similar gaps have shown up in benchmarks for protein structure reasoning and chemical synthesis planning. MatPhaseBench adds a rigorous, human-validated datapoint to that growing list, and sets a bar that current frontier models do not clear.","[\"ai\",\"materials-science\",\"benchmarks\",\"vision-language-models\"]","2026-07-07T04:00:00.000Z","2026-07-07T08:27:22.780Z","2026-07-07T08:27:25.742Z","published",null,[],"ai",[24,26,27,28],"materials-science","benchmarks","vision-language-models",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.02934",0,{"sections":35},[36,40,45,50,55,60,65,70,75,79,84,88,93,98],{"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":18},"Dev Tools","dev-tools",59,{"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"]