:root{--paper:#f6f5f0;--ink:#203b2f;--muted:#687467;--green:#356546;--line:#d5dbce;--soft:#e9ede2;--rust:#944b35;--rust-soft:#f2e7de;--serif:Georgia,'Times New Roman',serif;--sans:Inter,ui-sans-serif,system-ui,-apple-system,BlinkMacSystemFont,'Segoe UI',sans-serif}*{box-sizing:border-box}html{scroll-behavior:smooth;scroll-padding-top:30px}body{margin:0;background:var(--paper);color:var(--ink);font-family:var(--sans);font-size:15px;-webkit-font-smoothing:antialiased}a{color:inherit;text-underline-offset:4px}a:hover{color:var(--green)}button{font:inherit;cursor:pointer}button:focus-visible,a:focus-visible,summary:focus-visible,[tabindex]:focus-visible{outline:3px solid #759745;outline-offset:4px}button:disabled{cursor:wait;opacity:.65}svg{fill:none;stroke:currentColor;stroke-width:1.6}.wrap{width:min(1260px,calc(100% - 80px));margin-inline:auto}.site-header{height:96px;display:flex;align-items:center;justify-content:space-between;border-bottom:1px solid var(--line);gap:24px}.brand{display:flex;align-items:center;gap:11px;font-family:var(--serif);font-size:23px;text-decoration:none;white-space:nowrap}.brand svg{height:33px;width:33px}.site-header nav{display:flex;gap:30px;font-size:12px}.site-header nav a{text-decoration:none}.site-header nav [aria-current]{border-bottom:1px solid var(--green);padding-bottom:5px}.skip-link{position:absolute;top:-80px;left:15px;padding:12px;background:var(--ink);color:white;z-index:10}.skip-link:focus{top:12px}.eyebrow{font-size:10px;font-weight:600;letter-spacing:1.5px;line-height:1.7;margin:0 0 20px;color:var(--muted)}.signal-dot{display:inline-block;width:7px;height:7px;border-radius:50%;background:var(--green);margin-right:9px}.hero{padding-block:72px 46px}.hero h1{font-family:var(--serif);font-size:clamp(55px,6.6vw,90px);line-height:1.04;letter-spacing:-3.5px;font-weight:400;margin:20px 0 28px}.hero h1 em{font-weight:400;color:var(--green)}.hero-subtitle{font-size:clamp(19px,2.2vw,28px);line-height:1.45;margin:0 0 22px;color:var(--green)}.deck{font-size:19px;line-height:1.75;max-width:735px;margin:0;color:#50624f}.alignment-visual{display:grid;grid-template-columns:1fr 36px 1fr 36px 1fr;gap:18px;align-items:center;margin-top:52px;padding:29px 33px;background:var(--soft);border:1px solid var(--line);border-radius:5px}.alignment-visual div{min-width:0}.alignment-visual span{font-size:9px;letter-spacing:1.3px;color:var(--muted)}.alignment-visual p{font-family:var(--serif);font-size:21px;line-height:1.4;margin:12px 0 0}.alignment-visual i{font-style:normal;font-size:25px;color:#8b9e83}.visual-caption{font-size:11px;color:var(--muted);line-height:1.7;margin:12px 0 0}.article-layout{display:grid;grid-template-columns:190px minmax(0,1fr);gap:70px;padding-top:48px;border-top:1px solid var(--line)}.contents{position:sticky;top:30px;align-self:start}.contents .eyebrow{font-size:9px}.contents nav{display:flex;flex-direction:column;gap:17px}.contents nav a{font-size:11px;text-decoration:none;line-height:1.5;color:var(--muted)}.contents nav a:hover{color:var(--ink)}.contents-gallery{display:block;font-size:10px;margin-top:34px;padding-top:20px;border-top:1px solid var(--line)}article{min-width:0}article>section{padding-bottom:62px;margin-bottom:58px;border-bottom:1px solid var(--line)}article>section:last-child{border:0;margin-bottom:0}.prose{max-width:720px}.prose p{font-size:15px;line-height:1.9;margin:19px 0;color:#40513f}.prose p.eyebrow{font-size:10px;line-height:1.7;margin:0 0 17px;color:var(--muted)}h2{font-family:var(--serif);font-size:clamp(30px,3.2vw,44px);line-height:1.16;font-weight:400;letter-spacing:-1.1px;margin:0 0 26px}h3{font-family:var(--serif);font-size:26px;line-height:1.3;letter-spacing:-.45px;font-weight:400;margin:26px 0 12px}strong{font-weight:600}code{font-family:ui-monospace,SFMono-Regular,Consolas,monospace;font-size:.84em;background:#e8eadf;padding:2px 4px;border-radius:3px;overflow-wrap:anywhere}.formula{background:var(--ink);color:var(--paper);padding:25px 28px;margin:26px 0 25px;border-radius:5px}.formula>span{display:block;letter-spacing:1px;text-transform:uppercase;font-size:10px;color:#bacbb1;margin-bottom:18px}.formula strong{font-family:var(--serif);font-size:25px;display:flex;align-items:center;gap:15px;font-weight:400}.fraction{display:inline-grid;text-align:center;font-family:var(--sans);font-size:12px;line-height:1.7;gap:7px}.fraction>span:first-child{border-bottom:1px solid #93a48a;padding:0 7px 7px}.scope-control{display:flex;gap:4px;border:1px solid var(--line);border-radius:5px;width:fit-content;padding:4px;margin-top:28px}.scope-control button{background:transparent;color:var(--muted);border:0;border-radius:3px;padding:10px 15px;font-size:11px}.scope-control button[aria-pressed=true]{background:var(--ink);color:var(--paper)}.small{font-size:11px;line-height:1.8;color:var(--muted);margin:14px 0 20px}.comparison{display:grid;grid-template-columns:1fr 1fr;gap:18px;margin:22px 0}.score-card{border:1px solid var(--line);border-radius:5px;padding:23px 20px 15px;min-width:0;background:#faf9f5}.score-card.after{background:#edf1e6;border-color:#c4d0ba}.score-top{display:flex;align-items:center;gap:10px;justify-content:space-between;font-size:10px;color:var(--muted)}.score-top h3{font-family:var(--sans);font-size:12px;margin:0;letter-spacing:0;font-weight:600;color:var(--ink)}.score-value{font-family:var(--serif);font-size:53px;letter-spacing:-2px;margin:18px 0 1px;line-height:1.2}.score-value span{font-size:29px}.score-label{font-size:10px;letter-spacing:.2px;color:var(--muted);margin:0 0 14px}.score-equation{font-size:11px;font-variant-numeric:tabular-nums;padding-bottom:18px;border-bottom:1px solid var(--line);margin:0}.cross-table{border-collapse:collapse;width:100%;font-size:11px;margin:18px 0 15px;font-variant-numeric:tabular-nums}.cross-table caption{text-align:left;font-size:10px;color:var(--muted);padding-bottom:13px}.cross-table th,.cross-table td{padding:11px 7px;border-bottom:1px solid var(--line);text-align:right}.cross-table th{font-weight:500;font-size:10px}.cross-table th:first-child{text-align:left;padding-left:0;line-height:1.5}.cross-table th small{display:block;font-weight:400;color:var(--muted);font-size:9px}.cross-table .mismatch{background:var(--rust-soft);color:var(--rust);font-weight:700;border-radius:3px}.cross-table .unknown{color:var(--muted)}.native-score{font-size:11px;line-height:1.7;margin:14px 0 0}.native-score strong{font-size:13px}.coverage{font-size:10px;line-height:1.7;color:var(--muted);margin:10px 0 0}.callout{border-left:3px solid #7c946b;background:var(--soft);padding:18px 22px;margin:26px 0;font-size:12px;line-height:1.8;max-width:800px}.callout>strong{font-size:13px}.callout p{margin:7px 0 0;color:#50624f}.callout.correction{border-color:#b57c5e;background:#f0eae0}.lineage{display:flex;gap:3px;margin-top:24px;border-radius:4px;overflow:hidden}.lineage span{background:#dce3d3;font-size:10px;padding:14px 10px;white-space:nowrap;text-align:center;min-width:91px}.lineage span:nth-child(2){background:#a7bc92}.lineage span:nth-child(3){background:#476a48;color:white}.table-scroll{overflow:auto;border:1px solid var(--line);border-radius:5px;margin:25px 0}.instruction-table{border-collapse:collapse;width:100%;min-width:690px;font-size:11px;line-height:1.7}.instruction-table th{font-size:10px;letter-spacing:.2px;text-align:left;padding:13px;background:var(--soft);font-weight:600}.instruction-table td{vertical-align:top;padding:16px 13px;border-top:1px solid var(--line)}.instruction-table th:first-child{width:105px}.instruction-table th:last-child{width:115px}.instruction-table td:last-child{font-variant-numeric:tabular-nums;white-space:nowrap}.instruction-table td:nth-child(2){color:var(--muted);width:30%}.task-ref{font-size:10px;line-height:1.7}.task-ref span{display:block;white-space:nowrap}.delta{display:block;color:var(--green);font-size:10px;margin-top:5px}.delta.negative{color:var(--rust)}.revision-label{display:block;font-size:9px;color:var(--muted);margin-top:7px}.case{border:1px solid var(--line);border-radius:5px;margin:28px 0 36px;overflow:hidden;background:#faf9f5}.case-head{padding:24px 25px 19px;display:flex;gap:20px;justify-content:space-between;align-items:flex-start}.case-head .eyebrow{font-size:9px;margin-bottom:9px}.case-head h3{margin:0;font-size:27px}.task-rate{text-align:right;white-space:nowrap;font-size:10px;color:var(--muted)}.task-rate strong{display:block;font-family:var(--serif);font-size:25px;color:var(--green);font-weight:400;line-height:1.4}.task-rate.negative strong{color:var(--rust)}.paired-videos{display:grid;grid-template-columns:1fr 1fr;gap:1px;background:var(--line);border-block:1px solid var(--line)}.clip{margin:0;background:#faf9f5;min-width:0}.clip-header{display:flex;align-items:center;justify-content:space-between;gap:8px;padding:11px 14px;font-size:10px}.status{font-size:9px;white-space:nowrap;border-radius:3px;background:#e1eada;color:var(--green);padding:3px 6px}.status.failure{background:var(--rust-soft);color:var(--rust)}video{display:block;width:100%;height:auto;aspect-ratio:2/1;background:#17241c}.clip figcaption{padding:13px 15px 16px}.clip-instruction{font-size:11px;line-height:1.7;margin:0 0 10px}.clip-meta{font-size:9px;color:var(--muted);line-height:1.7;margin:0}.clip-meta a{white-space:nowrap}.case-foot{padding:18px 25px 23px}.case-foot>p{font-size:12px;line-height:1.8;margin:0 0 13px;color:#50624f}.case-foot .case-detail{font-size:12px;color:#40513f}.case-actions{display:flex;gap:12px;align-items:center;flex-wrap:wrap}.play-pair{border:1px solid var(--line);border-radius:3px;background:var(--soft);color:var(--ink);padding:9px 12px;font-size:10px}.case-source{font-size:10px;color:var(--muted)}.play-status{font-size:10px;color:var(--rust);margin:0}.training-flow{display:flex;flex-direction:column;align-items:center;gap:6px;margin:25px 0;padding:25px;background:var(--soft);border-radius:5px}.training-flow span{font-size:12px;border:1px solid #c5d1bc;background:var(--paper);padding:10px 18px;width:min(310px,100%);text-align:center}.training-flow b{font-weight:400;color:#859879}.failure-grid{display:grid;grid-template-columns:1fr 1fr;gap:20px;margin:27px 0}.failure-case{border:1px solid var(--line);border-radius:5px;overflow:hidden;background:#faf9f5}.failure-case h3{font-size:23px;margin:0 0 12px}.failure-description{padding:20px}.failure-description>.eyebrow{font-size:9px;margin-bottom:10px}.failure-description>p:not(.eyebrow){font-size:12px;line-height:1.8;color:#50624f;margin:0}.failure-case .clip-instruction{font-size:10px}.conclusion{border-left:3px solid var(--green);padding-left:25px!important}.methods h2{font-size:34px}.methods details{border-bottom:1px solid var(--line);padding:17px 0}.methods summary{cursor:pointer;font-size:12px;font-weight:600;line-height:1.6}.methods details p{font-size:12px;line-height:1.85;margin:15px 0 0}.explore{border-top:1px solid var(--line);padding-block:43px 45px;display:flex;justify-content:space-between;align-items:center;gap:30px}.explore h2{font-size:33px;margin:0 0 12px}.explore .eyebrow{margin-bottom:15px}.explore p:not(.eyebrow){font-size:12px;color:var(--muted);line-height:1.7;margin:0}.explore nav{display:grid;grid-template-columns:1fr 1fr;gap:12px}.explore nav a{border:1px solid var(--line);padding:14px 20px;min-width:145px;text-decoration:none;font-size:12px}.explore nav a:hover{background:var(--soft)}.site-footer{border-top:1px solid var(--line);padding-block:25px;display:flex;justify-content:space-between;gap:20px;font-size:10px;color:var(--muted)}.noscript{padding:30px;text-align:center;font-size:12px;line-height:1.8}#load-error{background:var(--rust-soft);color:var(--rust);padding:20px}[hidden]{display:none!important}
@media(min-width:1550px){.wrap{width:min(1360px,calc(100% - 120px))}.article-layout{grid-template-columns:200px minmax(0,1fr);gap:100px}.hero{padding-top:90px}.hero h1{font-size:96px}}
@media(max-width:1050px){.wrap{width:calc(100% - 52px)}.article-layout{grid-template-columns:145px minmax(0,1fr);gap:30px}.hero h1{font-size:70px}.alignment-visual{padding:25px;gap:12px}.alignment-visual p{font-size:18px}.score-card{padding:20px 15px 15px}.cross-table{font-size:10px}.case-head{padding:20px}.case-head h3{font-size:25px}.task-rate strong{font-size:22px}.case-foot{padding:17px 20px}.failure-grid{gap:14px}}
@media(max-width:800px){.contents{display:none}.article-layout{display:block}.hero{padding-top:48px}.hero h1{font-size:65px}.deck{font-size:17px}.alignment-visual{grid-template-columns:1fr 22px 1fr 22px 1fr;gap:9px;padding:22px 20px}.alignment-visual p{font-size:17px}.alignment-visual span{font-size:8px;letter-spacing:.8px}.alignment-visual i{font-size:19px}.prose{max-width:none}.explore{align-items:flex-start}.explore nav{gap:8px}.explore nav a{padding:12px;min-width:122px}}
@media(max-width:560px){.wrap{width:calc(100% - 36px)}.site-header{height:75px;gap:10px}.brand{font-size:19px;gap:7px}.brand svg{height:27px;width:27px}.site-header nav{gap:14px;font-size:10px}.site-header nav a:last-child{display:none}.hero{padding-block:37px 32px}.hero .eyebrow{font-size:8px;letter-spacing:1px}.hero h1{font-size:49px;letter-spacing:-2px;line-height:1.07;margin:20px 0}.deck{font-size:15px;line-height:1.8}.alignment-visual{grid-template-columns:1fr;gap:8px;margin-top:30px;padding:21px 23px}.alignment-visual span{font-size:8px}.alignment-visual p{margin:6px 0 0;font-size:18px}.alignment-visual i{font-size:18px;transform:rotate(90deg);width:16px}.visual-caption{font-size:10px}.article-layout{padding-top:33px}article>section{padding-bottom:36px;margin-bottom:35px}.prose p{font-size:14px;line-height:1.85}.prose p.eyebrow{font-size:8px;letter-spacing:1.1px;margin-bottom:14px}h2{font-size:32px;letter-spacing:-.85px;margin-bottom:22px}h3{font-size:25px}.formula{padding:20px 17px}.formula strong{font-size:24px;gap:10px}.formula>span{font-size:9px}.fraction{font-size:9px;line-height:1.6}.scope-control{width:100%}.scope-control button{padding:10px 7px;font-size:10px;flex:1}.small{font-size:10px}.comparison{grid-template-columns:1fr;gap:13px}.score-card{padding:21px}.score-value{font-size:56px;margin-top:15px}.cross-table{font-size:12px}.cross-table th{font-size:11px}.cross-table th small{font-size:10px}.score-top h3{font-size:13px}.score-top{font-size:11px}.native-score{font-size:12px}.coverage{font-size:11px}.callout{padding:16px 18px;font-size:11px}.callout>strong{font-size:12px}.lineage span{min-width:75px;font-size:9px;padding:12px 7px}.instruction-table{min-width:660px}.case{margin:23px 0 29px}.case-head{padding:19px 17px;gap:10px}.case-head h3{font-size:24px}.case-head .eyebrow{font-size:8px}.task-rate{font-size:8px}.task-rate strong{font-size:20px}.paired-videos{grid-template-columns:1fr}.clip-header{font-size:11px;padding:11px 14px}.clip-instruction{font-size:12px}.clip-meta{font-size:10px}.clip figcaption{padding:13px 15px 17px}.case-foot{padding:18px 17px}.case-foot>p,.case-foot .case-detail{font-size:12px}.play-pair{font-size:11px}.failure-grid{grid-template-columns:1fr;gap:22px}.failure-case .clip-instruction{font-size:11px}.failure-description{padding:18px}.conclusion{padding-left:19px!important}.conclusion h2{font-size:30px}.methods h2{font-size:30px}.methods summary{font-size:12px}.methods details p{font-size:12px}.explore{display:block;padding-block:32px}.explore nav{margin-top:25px}.explore nav a{font-size:12px;padding:14px}.site-footer{flex-direction:column;gap:14px;font-size:10px}}
@media(prefers-reduced-motion:reduce){html{scroll-behavior:auto}}

/* The operator test compares the agent's judgment with the benchmark's verdict. */
.study-flow{display:grid;grid-template-columns:repeat(4,minmax(0,1fr));margin-top:42px;border:1px solid var(--line);border-radius:5px;overflow:hidden;background:#edf0e6}.study-step{padding:24px 23px;border-right:1px solid var(--line);min-width:0}.study-step:last-child{border-right:0}.study-step>span{font-size:9px;letter-spacing:1.15px;color:var(--muted)}.study-step h3{font-size:23px;line-height:1.25;margin:14px 0 13px}.study-step p{font-size:11px;line-height:1.8;color:#50624f;margin:0}.comparison-step{background:#e0e8d5}.review-rule{display:grid;grid-template-columns:1fr 1fr;border:1px solid var(--line);border-radius:5px;background:var(--soft);margin:24px 0 22px;overflow:hidden}.review-rule>div{padding:23px}.review-rule>div+div{border-left:1px solid var(--line);background:#f0ece3}.review-rule .eyebrow{display:block;font-size:9px;letter-spacing:1px;margin:0 0 12px}.review-rule p{font-size:13px;line-height:1.8;margin:0 0 15px;color:#40513f}.review-rule strong{font-size:11px;line-height:1.7;display:block;color:var(--ink)}.metric-intro{font-size:15px;line-height:1.9;color:#40513f;margin-top:24px}
@media(max-width:1050px){.study-step{padding:21px 18px}.study-step h3{font-size:21px}.study-step>span{font-size:8px;letter-spacing:.8px}}
@media(max-width:800px){.study-flow{grid-template-columns:repeat(2,minmax(0,1fr))}.study-step:nth-child(2){border-right:0}.study-step:nth-child(n+3){border-top:1px solid var(--line)}.study-step h3{font-size:23px}.study-step p{font-size:12px}}
@media(max-width:560px){.study-flow{margin-top:29px}.study-step{padding:18px 15px}.study-step>span{font-size:7px;letter-spacing:.7px}.study-step h3{font-size:20px;line-height:1.3;margin:11px 0}.study-step p{font-size:10px;line-height:1.8}.review-rule{grid-template-columns:1fr}.review-rule>div{padding:19px}.review-rule>div+div{border-left:0;border-top:1px solid var(--line)}.review-rule p{font-size:12px}.review-rule strong{font-size:11px}.metric-intro{font-size:14px}}

/* Keep longer judgment labels readable in the comparison tables. */
.cross-table th:first-child{width:44%}.cross-table thead th:not(:first-child){width:28%}.instruction-table td strong{font-weight:750;color:var(--ink)}
@media(max-width:560px){.hero h1{font-size:44px}.fraction{font-size:10px}.cross-table th{font-size:10px}.cross-table th:first-child{width:42%}}

/* RoboTwin benchmark section. */
.robotwin-summary-grid{display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:12px;margin:18px 0 26px;max-width:800px}.mini-stat{border:1px solid var(--line);border-radius:5px;background:#faf9f5;padding:16px 15px}.mini-stat span{display:block;font-size:9px;letter-spacing:.7px;color:var(--muted);text-transform:uppercase;line-height:1.5}.mini-stat strong{display:block;font-family:var(--serif);font-size:30px;line-height:1.25;margin-top:8px;font-weight:400;color:var(--green)}.robotwin-cases{display:grid;grid-template-columns:1fr 1fr;gap:20px;margin:27px 0}.robotwin-case{border:1px solid var(--line);border-radius:5px;overflow:hidden;background:#faf9f5}.robotwin-case .case-head{display:block}.robotwin-case .case-head h3{font-size:23px}.robotwin-case .task-rate{text-align:left;margin-top:13px}.robotwin-case-body>p{font-size:12px;line-height:1.8;color:#50624f;margin:0;padding:18px 20px 20px}.robotwin-case .clip-instruction{font-size:10px}@media(max-width:800px){.robotwin-summary-grid{grid-template-columns:repeat(2,minmax(0,1fr))}.robotwin-cases{grid-template-columns:1fr}}@media(max-width:560px){.robotwin-summary-grid{grid-template-columns:1fr}.mini-stat strong{font-size:28px}}

/* Five-part essay: benchmark subsections and inspectable instruction repairs. */
.overview-film{max-width:1100px;margin:40px auto 0;background:#faf9f5;border:1px solid var(--line);border-radius:8px;overflow:hidden;box-shadow:0 12px 38px #203b2f08;scroll-margin-top:24px}.overview-heading{display:flex;align-items:center;justify-content:space-between;gap:20px;padding:23px 27px 22px}.overview-heading .eyebrow{font-size:9px;margin:0 0 8px;letter-spacing:1.4px}.overview-heading h2{font-size:29px;line-height:1.2;letter-spacing:-.5px;margin:0}.overview-duration{font-size:11px;color:var(--muted);white-space:nowrap;border:1px solid var(--line);border-radius:30px;padding:8px 13px}.overview-player{background:#17241c;border-block:1px solid var(--line)}.overview-player video{aspect-ratio:16/9;object-fit:contain;width:100%;height:auto}.overview-caption{display:flex;align-items:center;justify-content:space-between;gap:18px;padding:17px 27px;font-size:11px;line-height:1.8;color:var(--muted)}.overview-caption a{color:var(--ink);white-space:nowrap}#methods .study-flow{margin-top:30px}#methods .visual-caption{margin-bottom:26px}
@media(max-width:560px){.overview-film{margin-top:28px;border-radius:6px}.overview-heading{padding:17px 16px;gap:10px}.overview-heading h2{font-size:22px}.overview-heading .eyebrow{font-size:8px;margin-bottom:7px}.overview-duration{font-size:9px;padding:6px 9px}.overview-caption{align-items:flex-start;flex-direction:column;gap:8px;padding:13px 16px;font-size:10px}}
.key-takeaways{margin-top:34px;padding:27px 30px 30px;background:var(--soft);border:1px solid var(--line);border-left:3px solid var(--green);border-radius:5px}.key-takeaways h2{font-family:var(--sans);color:var(--green);margin:0 0 22px}.key-takeaways ol{display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:36px;margin:0;padding-left:20px}.key-takeaways li{padding-left:4px;min-width:0}.key-takeaways li::marker{font-size:12px;font-weight:600;color:var(--green)}.key-takeaways h3{font-size:23px;line-height:1.3;margin:0 0 13px}.key-takeaways p{font-size:13px;line-height:1.8;margin:0;color:#40513f}
@media(max-width:800px){.key-takeaways ol{grid-template-columns:1fr;gap:22px}.key-takeaways h3{font-size:24px;margin-bottom:9px}}
@media(max-width:560px){.key-takeaways{margin-top:26px;padding:21px 20px 23px}.key-takeaways h3{font-size:22px}.key-takeaways p{font-size:12px}.key-takeaways h2{margin-bottom:19px}}
.hero h1{font-size:clamp(36px,5.8vw,78px);max-width:1160px;letter-spacing:-2.5px}.anchor-alias{display:block;scroll-margin-top:30px}.benchmark-section{margin-top:35px}.benchmark-section+.benchmark-section{border-top:1px solid var(--line);padding-top:38px;margin-top:55px}.subsection-title{font-size:clamp(26px,3vw,35px);margin:0 0 24px}.prose h4{font-size:18px;line-height:1.5;font-weight:600;margin:28px 0 12px;letter-spacing:-.2px}.benchmark-table{border-collapse:collapse;width:100%;min-width:640px;font-size:12px;line-height:1.8}.benchmark-table th,.benchmark-table td{padding:16px;text-align:left;vertical-align:top;border-bottom:1px solid var(--line)}.benchmark-table thead th{background:var(--soft);font-size:11px}.benchmark-table tbody th{white-space:nowrap}.failure-patterns{display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:16px;margin:25px 0}.failure-patterns>div{padding:20px;border:1px solid var(--line);border-radius:5px;background:#edf0e6}.failure-patterns h3{font-size:22px;margin:0 0 13px}.failure-patterns p{font-size:12px;line-height:1.85;margin:0;color:#40513f}.instruction-comparison{border-top:1px solid var(--line);margin-top:14px;padding-top:13px}.instruction-label{display:block;font-size:10px;font-weight:600;letter-spacing:.2px;margin:0 0 6px;color:var(--ink)}.instruction-base,.instruction-goal-criteria,.instruction-full,.instruction-baseline{white-space:pre-wrap;overflow-wrap:anywhere;font-size:11px;line-height:1.8;margin:6px 0 12px;color:#40513f}.instruction-comparison details{border-top:1px solid var(--line);padding:10px 0}.instruction-comparison summary{font-size:11px;font-weight:600;line-height:1.7;cursor:pointer}.instruction-provenance{font-size:10px;line-height:1.8;color:var(--muted);margin:10px 0 0;overflow-wrap:anywhere}.clip-header{flex-wrap:wrap}.clip-header>span:first-child{line-height:1.6}.robotwin-case .task-rate strong{font-size:22px;white-space:normal}.methods{margin-top:26px}.methods details p{overflow-wrap:anywhere}
.robotwin-case .instruction-comparison{padding:16px 20px 20px}
.failure-description .task-rate{text-align:left;margin:14px 0}.failure-description .task-rate strong{font-size:28px;line-height:1.4}#failures .prose p.small{font-size:11px;line-height:1.8;color:var(--muted)}
@media(max-width:1050px){.failure-patterns{grid-template-columns:1fr}.failure-patterns>div{padding:19px 22px}.failure-patterns h3{font-size:23px}}
@media(max-width:560px){.hero h1{font-size:38px;letter-spacing:-1.4px}.benchmark-section+.benchmark-section{margin-top:35px;padding-top:30px}.instruction-base,.instruction-goal-criteria,.instruction-full,.instruction-baseline{font-size:11px}.instruction-provenance{font-size:10px}.failure-patterns{gap:12px}.subsection-title{font-size:27px}.benchmark-table{font-size:11px}}

/* Keep a bounded alignment score readable when completion judgments are unpublished. */
.score-value.score-range{font-size:clamp(30px,3.3vw,45px);letter-spacing:-1.5px}.cross-table .unclassified{background:#f0ece3}
