@import"https://fonts.googleapis.com/css2?family=EB+Garamond:wght@400;500;600;700&family=IBM+Plex+Mono:wght@400;500&display=swap";:root{--font-text: "EB Garamond", "Times New Roman", serif;font-family:var(--font-text);font-size:18px;color:#222;background:#fff;font-synthesis:none;font-weight:400;-webkit-font-smoothing:antialiased}*{box-sizing:border-box}body{margin:0}a{color:inherit;text-decoration:none}button,input,select{font:inherit}button,a,input,select,summary{outline-offset:4px}button{cursor:pointer}button:disabled{cursor:default}button,a,select{-webkit-tap-highlight-color:transparent}.site-header{height:82px;padding:0 5%;border-bottom:1px solid #dce1da;display:flex;align-items:center;gap:18px;background:#fff}.brand{display:flex;align-items:center;gap:10px;font-size:20px;letter-spacing:-.6px}.lab-brand{display:flex;align-items:center;gap:6px;font-size:11px;color:#222;border-left:1px solid #dce1da;padding-left:16px;white-space:nowrap}.lab-mark{flex:none;overflow:visible;color:#222;transition:transform .12s ease}.lab-mark-back,.lab-mark-top{transform-box:fill-box;transform-origin:center;transition:transform .32s cubic-bezier(.2,.75,.2,1)}.lab-mark-top{transform:rotate(9.80665deg)}.lab-brand:is(:hover,:focus-visible) .lab-mark-top{transform:translate(-3px,3px) rotate(0)}.lab-brand:is(:hover,:focus-visible) .lab-mark-back{transform:translate(2px,-2px)}.lab-brand:active .lab-mark{transform:scale(.9)}.lab-mark.is-entering .lab-mark-top{animation:lab-settle-front .82s cubic-bezier(.2,.75,.2,1) both}.lab-mark.is-entering .lab-mark-back{animation:lab-settle-back .82s cubic-bezier(.2,.75,.2,1) both}@keyframes lab-settle-front{0%{transform:translate(18px,-13px) rotate(-7deg)}to{transform:translate(0) rotate(9.80665deg)}}@keyframes lab-settle-back{0%{transform:translate(-11px,10px) rotate(-5deg)}to{transform:translate(0) rotate(0)}}@media(prefers-reduced-motion:reduce){.lab-mark,.lab-mark-back,.lab-mark-top,.lab-mark.is-entering .lab-mark-back,.lab-mark.is-entering .lab-mark-top{animation:none;transition:none}.lab-brand:is(:hover,:focus-visible) .lab-mark-top{transform:rotate(9.80665deg)}.lab-brand:is(:hover,:focus-visible) .lab-mark-back,.lab-brand:active .lab-mark{transform:none}}.site-header nav{margin-left:auto}.beta{font-size:11px;letter-spacing:0;border:1px solid #d3d9d1;padding:3px 6px;color:#737e73;font-weight:500;margin-left:2px}nav{display:flex;gap:34px;height:100%;align-items:center}nav a{font-size:14px;color:#6c746c;height:100%;display:flex;align-items:center}nav a[aria-current]{color:#1e3425;text-decoration:underline 2px #345c40;text-underline-offset:6px}.github-button{display:grid;place-items:center;width:28px;height:28px;margin-right:-20px;padding:0;border:0;background:none;cursor:pointer}.github-button img{display:block}.github-button:hover img{opacity:.75}.github-button:focus-visible{outline:2px solid #527352;outline-offset:2px}main{max-width:1134px;padding:40px 42px 0;margin:auto}.heading{display:flex;align-items:center;justify-content:space-between;gap:20px;margin-bottom:24px}h1{font-size:30px;letter-spacing:-1.1px;line-height:1.2;font-weight:600;margin:0 0 12px}h2{font-size:16px;font-weight:600;letter-spacing:-.2px;margin:0}p{font-size:16px;color:#5d6759;line-height:1.7;margin:0}.release{font:12px var(--font-text);color:#778174;white-space:nowrap}.table-scroll{position:relative;overflow:auto}table{border-collapse:collapse;width:100%;font-size:13px;text-align:left}th{font-weight:500;color:#7a8478;font-size:12px;border-bottom:1px solid #cfd7cc;padding:13px 14px;white-space:nowrap}td{border-bottom:1px solid #e0e5dd;padding:21px 14px;line-height:1.5}th:first-child,td:first-child{padding-left:0}th:last-child,td:last-child{padding-right:0}.leaderboard td{padding-top:18px;padding-bottom:18px}.leaderboard td:first-child{width:48px}.rank{color:#8c9588;font:12px var(--font-text)}.model-name{display:block;font-size:15px;font-weight:500;white-space:nowrap}.model-org,small{display:block;font-size:11px;color:#828c7d;margin-top:4px}.rate{min-width:170px;padding-right:38px}.rate strong{font-size:21px;font-weight:500;letter-spacing:-.7px;font-variant-numeric:tabular-nums}.rate strong span{font-size:12px;margin-left:1px;color:#6f7b6a}.bar{height:3px;background:#e4e9e1;max-width:116px;margin-top:8px}.bar i{height:100%;background:#477c54;display:block}.text-button{background:none;border:0;padding:3px 0;display:inline-flex;align-items:center;gap:5px;white-space:nowrap;font-size:12px;color:#4f6953}.text-button:hover,.task-link:hover,a:hover{color:#1c7c36}.muted{color:#899282}.disclosure{font-size:12px;line-height:1.8;margin-top:20px;color:#7f897b;max-width:950px}.section-line{display:flex;align-items:center;justify-content:space-between;margin:48px 0 16px}.section-line a{display:flex;align-items:center;gap:4px;font-size:12px;color:#71816b}.quick-tasks{border-top:1px solid #dbe2d7}.quick-tasks button{display:flex;justify-content:space-between;align-items:center;gap:20px;background:none;border:0;border-bottom:1px solid #e1e7dc;width:100%;text-align:left;font-size:13px;padding:15px 0;color:#596753}.quick-tasks button:hover{color:#1c7c36}.quick-tasks svg{flex-shrink:0}.filters{display:flex;align-items:center;gap:10px;margin-bottom:22px}.search{display:flex;align-items:center;gap:9px;border:1px solid #d8dfd3;background:#fff;max-width:400px;width:100%;padding:11px 12px;color:#8a9683}.search input{border:0;background:none;outline:none;width:100%;font-size:13px;color:#283423}.filters select{border:1px solid #d8dfd3;background:#fff;font-size:12px;padding:11px;color:#52604b}.task-cell{max-width:450px;min-width:220px}.task-link{background:none;border:0;padding:0;text-align:left;color:#34432e;font-size:13px;line-height:1.6}.task-cell small{overflow-wrap:anywhere;font:10px var(--font-text);max-width:390px;color:#939c8d}.outcome{font-size:12px;display:inline-flex;align-items:center;gap:6px}.outcome:before{content:"";width:5px;height:5px;background:currentColor}.pass{color:#3e7548}.fail{color:#a06a59}.withheld{color:#8c876a}code{font:11px IBM Plex Mono,monospace;color:#798770;overflow-wrap:anywhere}.more{border:1px solid #d5ded0;background:#fff;color:#687761;padding:10px 16px;display:block;margin:24px auto;font-size:12px}.empty{padding:40px}.env-list{display:grid;grid-template-columns:180px 1fr;border-top:1px solid #d5ddd0;font-size:13px;margin:0}.env-list dt,.env-list dd{margin:0;padding:18px 0;border-bottom:1px solid #e0e6db;line-height:1.6}.env-list dt{color:#7f8b76}.env-list dd{overflow-wrap:anywhere}details{margin-top:24px;border-top:1px solid #dbe2d7;padding-top:18px;font-size:13px}summary{cursor:pointer}pre{background:#f0f3ed;padding:20px;overflow:auto;font:11px/1.7 IBM Plex Mono,monospace;color:#67765e}footer{display:flex;justify-content:space-between;align-items:center;border-top:1px solid #dce3d7;padding:24px 0;margin-top:60px;font-size:11px;color:#8a9581}footer a{display:inline}.sr-only{position:absolute;width:1px;height:1px;overflow:hidden;clip:rect(0,0,0,0)}@media(max-width:700px){.site-header{height:66px;padding:0 20px}.brand{font-size:16px;gap:6px}.lab-brand{display:none}.beta{font-size:9px}nav{gap:18px}.github-button{margin-right:-10px}nav a{font-size:12px}main{padding:32px 20px 0}.heading{align-items:flex-start;flex-direction:column;gap:12px;margin-bottom:24px}h1{font-size:27px}.heading p{font-size:12px}.release{font-size:10px}td{padding:16px 12px}.leaderboard td{padding-top:22px;padding-bottom:22px}.leaderboard{min-width:830px}.rate{min-width:160px}.disclosure{font-size:11px}.section-line{margin-top:32px}.filters{flex-wrap:wrap}.search{max-width:none}.env-list{grid-template-columns:110px 1fr;gap:0 14px;font-size:12px}.env-list code{font-size:10px}footer{align-items:flex-start;gap:16px;line-height:1.6;margin-top:40px}.model-name{font-size:13px}}.model-identity{display:flex;align-items:center;gap:13px}.lab-logo{width:28px;height:28px;object-fit:contain;flex-shrink:0}.task-grid{display:grid;grid-template-columns:repeat(4,minmax(0,1fr));gap:16px}.task-card{display:flex;flex-direction:column;border:1px solid #dde3d9;color:inherit;transition:border-color .15s}.task-card:hover{border-color:#a8b5a0}.task-card img{display:block;width:100%;height:auto;aspect-ratio:1;border-bottom:1px solid #e6eae3;transition:filter .15s,opacity .15s}.task-card[data-locked=grey] img{filter:grayscale(1);opacity:.55}.task-card[data-locked=grey]:hover img{opacity:.8}.task-card[data-locked=silhouette] img{opacity:.85}.task-card[data-locked=silhouette]:hover img{opacity:1}.task-card h3{flex:1;margin:12px 14px 4px;font-size:15px;font-weight:500;line-height:1.35;letter-spacing:-.2px}.task-card-meta{display:flex;align-items:center;gap:5px;margin:0 14px 12px;font-size:12px;color:#5d6759}.task-card[data-locked] .task-card-meta{color:#8a9582}.compare{border:1px solid #dbe2d7;background:#fbfcf9;margin-bottom:28px}.compare-head{padding:16px 20px 14px;border-bottom:1px solid #dbe2d7}.compare-head h2{margin:0;font-size:20px;font-weight:500;letter-spacing:-.4px}.compare-head p{margin:4px 0 0;font-size:14px;color:#5d6759}.compare-pair{display:grid;grid-template-columns:minmax(0,1fr) 0px minmax(0,1fr);height:clamp(340px,calc(100dvh - 460px),540px);transition:grid-template-columns .35s ease}.compare-pair[data-revealed]{grid-template-columns:minmax(0,1fr) 320px minmax(0,1fr)}.compare-panel{display:flex;flex-direction:column;min-width:0;min-height:0;grid-row:1}.compare-panel[data-side=A]{grid-column:1}.compare-panel[data-side=B]{grid-column:3;border-left:1px solid #dbe2d7}.compare-pair[data-revealed] .compare-panel[data-side=B]{border-left:0}.compare-pair .sbs-checks{position:relative;grid-column:2;grid-row:1;max-height:none;overflow-x:hidden;border:0;border-left:1px solid #dbe2d7;border-right:1px solid #dbe2d7}.sbs-checks[data-reveal] .sbs-check,.sbs-checks[data-reveal] .sbs-scores{opacity:0;animation:check-in .28s ease-out forwards;animation-delay:calc(.4s + var(--i, 0) * 90ms)}.sbs-checks .sbs-scores{position:sticky;bottom:0;background:#fbfcf9;border-top:1px solid #dbe2d7;border-bottom:0}.sbs-checks .sbs-scores span:nth-child(2){color:#263329;font-size:12px;text-align:center}@keyframes check-in{0%{opacity:0;transform:translateY(6px)}to{opacity:1;transform:none}}@media(max-width:900px){.compare-pair,.compare-pair[data-revealed]{grid-template-columns:1fr;height:auto;transition:none}.compare-panel[data-side],.compare-pair .sbs-checks{grid-column:1}.compare-pair .sbs-checks{grid-row:2;border:0;border-top:1px solid #dbe2d7}.compare-panel[data-side=B]{grid-row:3;border-left:0;border-top:1px solid #dbe2d7}.sbs-checks .sbs-scores{position:static}.compare-panel .asset-preview,.compare-panel .sbs-placeholder{flex:none;aspect-ratio:1}}@media(prefers-reduced-motion:reduce){.compare-pair{transition:none}.sbs-checks[data-reveal] .sbs-check,.sbs-checks[data-reveal] .sbs-scores{animation:none;opacity:1}}.compare-panel header{display:flex;align-items:center;gap:7px;min-height:42px;padding:10px 12px;font-size:13px}.compare-panel header strong{font-weight:500}.compare-panel[data-picked] header{background:#eef2ea}.compare-panel .asset-preview{flex:1;min-height:0;border-width:1px 0 0}.compare-panel .sbs-placeholder{margin:0;border-width:1px 0 0}.compare-actions{display:flex;flex-wrap:wrap;justify-content:center;gap:8px;padding:14px;border-top:1px solid #dbe2d7}.compare-actions button{padding:8px 16px;border:1px solid #c3cbc0;background:#fff;font-size:14px;color:#263329}.compare-actions button:hover{background:#f0f2ed}.compare-actions .primary{background:#34482f;border-color:#34482f;color:#f0f4ed}.compare-actions .primary:hover{background:#2a3b26}.task-leaderboard{margin-top:36px}.task-leaderboard h2{font-size:18px;font-weight:500;letter-spacing:-.4px}.task-leaderboard .attempt-table{margin-top:10px}.attempt-table td.attempt-rank{width:56px;font:12px var(--font-text);color:#73796f}.turnstile-host:not(:empty){margin-top:24px}.task-leaderboard[data-locked] .attempt-table{mask-image:linear-gradient(#000 0 40%,transparent 100%)}.attempt-model-hidden{display:block;width:130px;height:14px;margin:7px 0;background:#e6eae3}.methodology-section{border-top:1px solid #dbe2d7;padding:28px 0}.methodology-section h2{font-size:21px;font-weight:500;letter-spacing:-.4px;margin-bottom:14px}.methodology-section>p{font-size:17px;line-height:1.65;max-width:750px}.methodology-joined{border-top:0;padding-top:0}.methodology-subtitle{margin-bottom:18px}.back-link{display:inline-flex;align-items:center;gap:6px;background:none;border:0;padding:0;font-size:13px;color:#5f7056;margin-bottom:24px}.back-link:hover{color:#1c7c36}.task-detail-header{margin-bottom:14px}.task-detail-header h1{margin:10px 0 0;font-size:32px;letter-spacing:-1.2px}.task-detail-brief{margin:20px 0 0;max-width:760px;font-size:18px;line-height:1.55;color:#2b3327;overflow-wrap:anywhere}.task-detail-brief span{display:block;margin-bottom:4px;font-size:11px;text-transform:uppercase;letter-spacing:.09em;color:#73796f}.task-detail-stats{display:flex;gap:20px;margin-top:10px;font:11px var(--font-text);color:#86917e}.task-detail-stats b{font-weight:500;color:#36482f;font-size:13px}.task-detail-models{border-top:1px solid #dbe2d7;margin-top:18px;padding-top:24px}.task-detail-models h2{font-size:18px;font-weight:500;letter-spacing:-.4px;margin-bottom:24px}.traces-table .model-identity{gap:10px}.traces-table .lab-logo{width:20px;height:20px}.traces-table .model-name{font-size:13px}.traces-table small{margin-top:2px}.traces-th-num{width:40px}.traces-rank{font:11px var(--font-text);color:#8c9588;width:40px}@media(max-width:960px){.task-grid{grid-template-columns:repeat(3,minmax(0,1fr))}}@media(max-width:700px){.task-grid{grid-template-columns:repeat(2,minmax(0,1fr));gap:10px}.task-card h3{margin:10px 10px 4px;font-size:14px}.task-card-meta{margin:0 10px 10px}.compare-head{padding:14px 16px}.compare-actions{display:grid;grid-template-columns:1fr 1fr}.task-detail-header h1{font-size:26px}.task-detail-brief{font-size:16px}.traces-th-num,.traces-rank{display:none}.traces-table .model-identity{gap:7px}.traces-table .lab-logo{width:16px;height:16px}}.task-detail-stats{font:13px var(--font-text);color:#5d6759}.task-detail-stats b{font-family:var(--font-text)}.task-detail-models h2{margin-bottom:6px}.task-detail-note{font-size:13px;color:#6b7466;margin-bottom:20px}.methodology-section>p+p{margin-top:12px}.check-families{display:grid;grid-template-columns:repeat(2,minmax(0,1fr));gap:0 40px;margin-top:22px;border-top:1px solid #dde3d9}.check-family{padding:18px 0 20px;border-bottom:1px solid #e6eae3}.check-family-head{display:flex;justify-content:space-between;align-items:baseline;gap:12px;margin-bottom:8px}.check-family-head h3{margin:0;font-size:18px;font-weight:500;color:#222}.check-family-head span{font-size:14px;color:#777;white-space:nowrap}.check-family-head b{font-weight:500;font-variant-numeric:tabular-nums}.check-family p{font-size:16px;line-height:1.6;color:#555}.check-family .check-family-example{margin-top:6px;font-size:15px;color:#777}.methodology-split{display:grid;grid-template-columns:minmax(0,1fr) 420px;gap:48px;align-items:start}.methodology-text>p{font-size:17px;line-height:1.65;max-width:750px}.methodology-text>p+p{margin-top:12px}.scoring-flow{display:grid;grid-template-columns:repeat(4,minmax(0,1fr));gap:0 34px;margin:6px 0 26px}.flow-stage{position:relative;min-width:0;container-type:inline-size}.flow-stage+.flow-stage:before{content:"→";position:absolute;left:-26px;top:50cqw;transform:translateY(-50%);color:#9aa596;font-size:17px}.flow-grid{display:grid;grid-template-columns:repeat(10,minmax(0,1fr));gap:3px;aspect-ratio:1}.flow-task{background:#dfe5db;border-radius:1px}.flow-brief{display:grid;place-items:center;background:#eef1ec}.flow-brief img{display:block;width:86%;height:86%;object-fit:contain;opacity:.8}.flow-split{display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:1px;background:none}.flow-split i{position:relative;background:#eef1ec;border-radius:1px;overflow:hidden}.flow-split b{position:absolute;inset:0;background:#2e3a2b;opacity:.8;mask-repeat:no-repeat;-webkit-mask-repeat:no-repeat}.flow-split i[data-outcome] b{background:#fff;opacity:.85}.flow-split i[data-outcome=pass]{background:#4c7a4f}.flow-split i[data-outcome=fail]{background:#c88a7e}.flow-split i[data-outcome=withheld]{background:#c9c4b5}.flow-board{list-style:none;margin:0;padding:0;aspect-ratio:1;display:grid;grid-auto-rows:minmax(0,1fr);align-items:center}.flow-board li{display:grid;grid-template-columns:14px minmax(0,1fr) 40px;align-items:center;gap:7px;opacity:.45}.flow-board li[data-shown]{opacity:1}.flow-board img{display:block}.flow-board span{height:7px;background:#eef1ec}.flow-board span i{display:block;height:100%;background:#8aa287}.flow-board li[data-shown] span i{background:#3f7a4d}.flow-board small{font:12px var(--font-text);color:#4f5a4c;text-align:right;font-variant-numeric:tabular-nums}.flow-stage p{margin:12px 0 0;font-size:13px;line-height:1.45;color:#777}.flow-stage p b{display:block;font-size:15px;font-weight:500;color:#222}.flow-pass{color:#4c7a4f}.flow-fail{color:#a3483a}@media(max-width:700px){.scoring-flow{grid-template-columns:repeat(2,minmax(0,1fr));gap:26px 22px}.flow-stage+.flow-stage:before{display:none}}@media(max-width:960px){.methodology-split{grid-template-columns:minmax(0,1fr);gap:24px}}.methodology-contact{border-top:1px solid #dbe2d7;padding-top:20px;font-size:14px;line-height:1.6;color:#777}.methodology-list{margin:14px 0;padding-left:22px;max-width:750px;font-size:17px;line-height:1.65;color:#555}.methodology-list li+li{margin-top:6px}.pipeline{display:grid;grid-template-columns:minmax(0,1fr) 400px;grid-template-areas:"heading figure" "steps figure";grid-template-rows:auto 1fr;gap:14px 48px;align-items:start}.pipeline>h2{grid-area:heading;margin:0}.pipeline-steps{grid-area:steps;margin:0}.pipeline-steps li{transition:color .16s ease}.pipeline-steps li+li{margin-top:12px}.pipeline-steps li[data-active]{color:#222}.pipeline-steps li[data-active]::marker{color:#34482f}.pipeline-figure{grid-area:figure;align-self:center;border:1px solid #dbe2d7;background:#fbfcfa}.pipeline-rail{display:grid;grid-template-columns:repeat(4,minmax(0,1fr));border-bottom:1px solid #dbe2d7}.pipeline-rail button{position:relative;padding:9px 0 10px;border:0;background:none;font-size:14px;color:#73796f}.pipeline-rail button:hover,.pipeline-rail button[aria-pressed=true]{color:#1e3425}.pipeline-rail span{position:absolute;left:0;right:0;bottom:-1px;height:2px;background:#34482f;transform-origin:left}.pipeline-stage{position:relative;aspect-ratio:400 / 320}.pipeline-stage>*{position:absolute;inset:0;display:block;width:100%;height:100%}.pipeline-stage[data-inspect]>svg,.pipeline-stage[data-inspect]>canvas{visibility:hidden}.pipeline-stage>.asset-preview{position:absolute;inset:0;width:100%;height:100%;aspect-ratio:auto;border:0;background:#fbfcfa}.pipeline-stage>.pipeline-inspect{inset:8px 12px auto auto;width:auto;height:auto;display:grid;place-items:center;padding:3px;border:0;background:none;color:#73796f}.pipeline-stage>.pipeline-inspect:hover{color:#263329}.pipeline-stage>.pipeline-inspect:focus-visible{outline:2px solid #527352;outline-offset:0}.pipeline-stage text{font-family:var(--font-text)}.pipeline-stage .pf-label{font-size:10.5px;letter-spacing:.09em;text-transform:uppercase;fill:#73796f}.pipeline-stage .pf-brief{font-size:15px;font-style:italic;fill:#2b3327}.pipeline-stage .pf-code{font:10.5px IBM Plex Mono,monospace;fill:#2b3327}.pf-caret{fill:#527352}.pf-code-bg{fill:#f0f2ed;stroke:#dbe2d7}.pf-slot{fill:none;stroke:#cfd6cb;stroke-dasharray:2 3}.pf-leader{fill:none;stroke:#6b7466}.pf-chip{fill:#fff;stroke-width:1.5}.pf-measure{fill:none;stroke:#183d27;stroke-width:1.6;stroke-linecap:round}.pipeline-stage .pf-note{font-size:13px;font-style:italic;fill:#183d27;paint-order:stroke;stroke:#fbfcfa;stroke-width:3px;stroke-linejoin:round}.pf-rule{stroke:#e2e6df}.pipeline-stage .pf-row{font-size:14px;fill:#263329}.pipeline-stage .pf-pass{font-size:13px;fill:#527352}.pipeline-stage .pf-fail{font-size:13px;fill:#97614f}.pipeline-stage .pf-fail-note{fill:#97614f}.pipeline-stage .pf-pass-note{fill:#527352}.pipeline-stage .pf-mark{font-style:normal}.pipeline-stage .pf-strike{stroke:#97614f;stroke-width:1.4;stroke-linecap:round}.pf-failed-row{fill:#97614f;fill-opacity:.07}.pf-fail-leader{stroke:#97614f}.pf-fail-dot{fill:#97614f}.pf-verdict-rule{stroke:#34482f;stroke-width:1.2}.pipeline-stage .pf-verdict{font-size:15px;font-weight:500;fill:#263329}.pipeline-stage .pf-verdict.pf-pass{fill:#527352}.pipeline-stage .pf-verdict.pf-fail{fill:#97614f}@media(max-width:960px){.pipeline{grid-template-columns:minmax(0,1fr);grid-template-areas:"heading" "figure" "steps";grid-template-rows:auto;gap:24px}.pipeline-figure{width:100%;max-width:480px}}@media(prefers-reduced-motion:reduce){.pipeline-steps li{transition:none}}@media(max-width:700px){.check-families{grid-template-columns:1fr}}@media(max-width:480px){.check-family-head{flex-wrap:wrap;gap:4px 12px}}.attempt-load{width:100%;height:100%;border:0;background:none;display:flex;flex-direction:column;align-items:center;justify-content:center;gap:4px;color:#4f5a4c;font-size:12px}.attempt-load small{font:11px var(--font-text);color:#6b7466}.attempt-load b{margin-top:6px;font-weight:500;font-size:12px;padding:5px 10px;border:1px solid #c3cbc0;background:#f6f7f3}.attempt-load:hover b{background:#fff}.sbs-compare{display:grid;grid-template-columns:minmax(0,1fr) 320px minmax(0,1fr);height:clamp(440px,calc(100dvh - 260px),680px);border:1px solid #dbe2d7;background:#fbfcf9;margin-top:8px}.sbs-panel{display:flex;flex-direction:column;min-width:0;min-height:0}.sbs-panel[data-side=left]{grid-column:1}.sbs-panel[data-side=right]{grid-column:3}.sbs-panel-head{display:grid;grid-template-columns:auto 1fr;row-gap:8px;column-gap:10px;align-items:center;padding:10px 12px;font-size:12px}.sbs-panel-head .sbs-model-name{grid-column:1 / -1}.sbs-panel[data-side=right] .sbs-rollout-pick{grid-column:2;grid-row:2;justify-self:end}.sbs-panel[data-side=right] .sbs-model-select{grid-column:1;grid-row:2}.sbs-model-name{display:flex;align-items:center;gap:7px;font-size:13px}.sbs-model-name strong{font-weight:500}.sbs-outcome{margin-left:auto;flex:none;font-size:11px}.sbs-outcome.pass{color:#3e7548}.sbs-outcome.fail,.sbs-outcome.withheld{color:#97614f}.sbs-rollout-pick{display:flex;align-items:center;gap:6px}.sbs-rollout-pick>span{font-size:10px;text-transform:uppercase;letter-spacing:.09em;color:#73796f}.sbs-seg{display:inline-flex;border:1px solid #c9d0c5;overflow:hidden}.sbs-seg button{min-width:30px;padding:4px 0;border:0;border-right:1px solid #c9d0c5;background:transparent;font:11px var(--font-text);cursor:pointer}.sbs-seg button:last-child{border-right:0}.sbs-seg button[data-outcome=pass]{color:#3e7548}.sbs-seg button[data-outcome=fail],.sbs-seg button[data-outcome=withheld]{color:#97614f}.sbs-seg button:hover{background:#f0f2ed}.sbs-seg button[aria-selected=true]{background:#e3e8df;color:#263329}.sbs-model-select{justify-self:end;max-width:100%;padding:4px 6px;border:1px solid #c9d0c5;background:#fff;font-size:11px;color:#263329}.sbs-panel[data-side=right] .sbs-model-select{justify-self:start}.sbs-panel .asset-preview{flex:1;min-height:0}.sbs-placeholder{flex:1;display:flex;flex-direction:column;align-items:center;justify-content:center;gap:4px;margin:0 12px 12px;border:1px solid #dfe2db;background:#f3f5f1;color:#6b7466;font-size:12px}.sbs-note{margin:0 12px 12px;font-size:11px;color:#6b7466}.sbs-placeholder[role=status]{padding:20px;text-align:center}.sbs-placeholder[role=status] small{max-width:100%;overflow-wrap:anywhere}.sbs-checks{grid-column:2;grid-row:1;min-width:0;min-height:0;overflow-y:auto;overscroll-behavior-y:contain;border-left:1px solid #dbe2d7;border-right:1px solid #dbe2d7;font-size:12px}.sbs-checks-header,.sbs-scores{display:grid;grid-template-columns:minmax(0,1fr) auto minmax(0,1fr);gap:8px;align-items:center}.sbs-check{display:grid;grid-template-columns:22px minmax(0,1fr) 22px;align-items:center}.sbs-checks-header{position:sticky;top:0;z-index:1;background:#fbfcf9;min-height:30px;padding:6px 10px 8px;border-bottom:1px solid #dbe2d7;color:#73796f;font:11px var(--font-text)}.sbs-checks-header span{white-space:nowrap;overflow:hidden;text-overflow:ellipsis}.sbs-checks-header strong{color:#263329;font:500 13px inherit;font-family:inherit;text-align:center}.sbs-checks-header span:last-child,.sbs-scores span:last-child,.sbs-check .sbs-state:last-child{text-align:right}.sbs-scores{padding:6px 10px;border-bottom:1px solid #dbe2d7;font:13px var(--font-text)}.sbs-scores [data-state=pass],.sbs-state[data-state=pass]{color:#3e7548}.sbs-scores [data-state=fail],.sbs-state[data-state=fail]{color:#a85f4c}.sbs-state[data-state=na]{color:#8a9086}.sbs-check{padding:6px 10px;border-bottom:1px solid #e5e7e1}button.sbs-check:hover{background:#f3f5f1}.sbs-state{font:11px var(--font-text)}.sbs-label{padding:0 8px;color:#3d473e}.attempt-table{width:100%;border-collapse:collapse;margin-top:20px;font-size:13px}.attempt-table th{text-align:left;font-weight:500;font-size:11px;color:#73796f;padding:8px 0;border-bottom:1px solid #dbe2d7}.attempt-table td{padding:7px 0;border-bottom:1px solid #e5e7e1;vertical-align:middle}.attempt-model{display:flex;align-items:center;gap:9px}.attempt-table td.model-score{font:12px var(--font-text)}.attempt-table td.model-score.pass{color:#3e7548}.attempt-table td.model-score.fail{color:#a06a59}.attempt-mark{display:inline-block;width:22px;font:12px var(--font-text)}.attempt-mark.mark-pass{color:#3e7548}.attempt-mark.mark-fail{color:#a85f4c}@media(max-width:900px){.sbs-compare{grid-template-columns:1fr;height:auto}.sbs-panel[data-side]{grid-column:1}.sbs-panel .asset-preview,.sbs-placeholder{aspect-ratio:1;flex:none}.sbs-checks{grid-column:1;grid-row:2;max-height:60dvh;border:0;border-top:1px solid #dbe2d7;border-bottom:1px solid #dbe2d7}.sbs-panel[data-side=right]{grid-row:3}}.sbs-check.sbs-metric-row{grid-template-columns:minmax(0,1fr) auto minmax(0,1fr);padding:10px;border-top:1px solid #dbe2d7}.sbs-metric{font:15px var(--font-text);color:#263329;font-variant-numeric:tabular-nums}.sbs-metric:last-child{text-align:right}button.sbs-check{position:relative;width:100%;border:0;border-bottom:1px solid #e5e7e1;background:none;font:inherit;color:inherit;text-align:inherit;cursor:pointer}button.sbs-check[aria-pressed=true]{background:#e6eddf}button.sbs-check[aria-pressed=true]:before{content:"";position:absolute;left:0;top:-1px;bottom:-1px;width:3px;background:#527352}button.sbs-check:focus-visible{outline:2px solid #527352;outline-offset:-2px}.bench-intro{display:grid;grid-template-columns:1.05fr 1fr;gap:60px;align-items:center;padding:10px 0 48px}.intro-byline{font:11px var(--font-text);color:#83907c;margin-bottom:14px}.bench-intro h1{font-size:56px;letter-spacing:-2.6px;font-weight:500;margin:0 0 18px;display:flex;align-items:center;gap:14px}.bench-intro h1 span{font:10px var(--font-text);border:1px solid #cfd8c9;color:#7a8872;padding:4px 7px;letter-spacing:0}.intro-description{font-size:19px;line-height:1.45;color:#34402e;margin-bottom:9px;letter-spacing:-.35px}.intro-explanation{font-size:13px;max-width:390px;line-height:1.75}.bench-facts{display:flex;gap:23px;font:10px var(--font-text);color:#86917e;margin-top:24px;white-space:nowrap}.bench-facts b{font-weight:500;color:#36482f;font-size:14px}.intro-assets{align-self:center}.intro-asset-strip{display:grid;grid-template-columns:repeat(3,1fr);gap:4px}.intro-asset-strip button{display:block;border:0;background:#e5e8e3;padding:0;position:relative;overflow:hidden}.intro-asset-strip img{width:100%;height:auto;display:block;transition:transform .25s}.intro-asset-strip button:hover img{transform:scale(1.04)}.intro-asset-strip svg{position:absolute;bottom:9px;right:9px;color:#78856f}.intro-assets>span{display:block;text-align:right;margin-top:10px;font-size:10px;color:#88947f}.home-section-heading{display:flex;align-items:center;justify-content:space-between;gap:14px;margin-bottom:24px}.home-section-heading h2{font-size:21px;font-weight:500;letter-spacing:-.6px}.home-section-heading>span,.home-section-heading>a{font-size:11px;color:#85917b}.home-section-heading>a{display:flex;align-items:center;gap:6px}.score-section{border-top:1px solid #d6dfd0;padding-top:28px}.scoreboard{position:relative}.score-head,.score-row,.score-axis{display:grid;grid-template-columns:240px minmax(0,1fr) 115px;align-items:center;gap:24px}.score-head{font-size:11px;color:#8a9682;padding:0 0 10px;border-bottom:1px solid #dbe2d6}.score-row{border:0;border-bottom:1px solid #e0e6dc;background:none;width:100%;text-align:left;padding:18px 0;color:inherit}.score-row:hover{background:#f0f4ed}.score-model{display:flex;align-items:center;gap:12px}.score-model img{object-fit:contain}.score-rank{width:22px;font:10px var(--font-text);color:#939d8a}.score-model strong{font-size:14px;font-weight:500}.score-model small{font-size:10px;margin-top:3px}.score-track{height:9px;background:#e9eee5;display:block;position:relative}.score-track>i{height:100%;display:block;background:#93a28c}.score-track .top-score{background:#477c54}.score-number{font-size:22px;letter-spacing:-.7px;font-variant-numeric:tabular-nums;display:flex;align-items:baseline;justify-content:flex-end;gap:2px}.score-number small{display:inline;font-size:11px;color:#78866e}.score-number svg{align-self:center;color:#9ba690;margin-left:15px}.score-axis{padding-top:10px}.score-axis>div{display:flex;justify-content:space-between;color:#9aa48f;font:9px var(--font-text)}.release-notes{display:flex;flex-wrap:wrap;align-items:baseline;justify-content:space-between;gap:12px;margin-top:22px;font-size:11px;color:#849179}.release-notes details{margin:0;border:0;padding:0;font-size:11px}.release-notes details[open]{flex-basis:100%;order:2}.release-notes details p{font-size:12px;margin:12px 0}.release-notes table{font-size:11px}.release-notes td{padding-top:10px;padding-bottom:10px}.issue-section,.example-section{padding-top:40px;margin-top:42px;border-top:1px solid #d6dfd0}.issue-grid{display:grid;grid-template-columns:repeat(4,minmax(0,1fr));gap:28px}.issue-frequency{display:block;margin-bottom:10px;white-space:nowrap}.issue-frequency strong{font-size:28px;font-weight:400;letter-spacing:-.8px}.issue-frequency>span{font-size:10px;color:#8a957e}.issue-item{display:flex;flex-direction:column;justify-content:flex-start;border:0;border-bottom:3px solid transparent;padding:12px 0 16px;background:none;color:inherit;font:inherit;text-align:left;cursor:pointer;border-radius:0}.issue-item:hover,.issue-item[aria-selected=true]{border-bottom-color:#527352}.issue-item:focus-visible{outline:2px solid #527352;outline-offset:4px}.issue-title{display:flex;align-items:center;justify-content:space-between;gap:8px;margin-bottom:15px;font-size:14px}.issue-description{display:block;font-size:13px;line-height:1.7;color:#5d6759}.failure-examples:not([hidden]){margin-top:28px}.failure-example-grid{display:grid;grid-template-columns:repeat(4,minmax(0,1fr));gap:24px}.failure-example .example-label{gap:8px;font-size:12px}.failure-example .example-model{min-width:0;color:inherit;text-decoration:none}.failure-example .example-model:hover{color:#527352}.failure-example .example-model>span{overflow-wrap:anywhere}.failure-example .example-fail{flex-shrink:0}.failure-example .asset-focus-status{display:none}@media(max-width:1100px){.failure-example-grid{grid-template-columns:repeat(2,minmax(0,1fr))}}@media(max-width:700px){.failure-example-grid{grid-template-columns:1fr;gap:28px}}.failure-pager{display:none}@media(max-width:700px){.failure-pager{display:flex;align-items:center;gap:6px;margin-bottom:12px;font-size:12px;color:#5d6759}.failure-pager span{margin-right:auto}.failure-pager button{display:grid;place-items:center;width:32px;height:32px;padding:0;border:1px solid #dfe2db;border-radius:0;background:#fff;color:#263329;cursor:pointer}.failure-pager button:active{background:#f0f3eb}.failure-pager button:focus-visible{outline:2px solid #527352;outline-offset:2px}.failure-example:not([data-current]){display:none}}.example-section>.home-section-heading{flex-wrap:wrap;gap:8px 24px}.example-tabs{display:flex;flex-shrink:0;gap:4px;margin-left:auto}.example-tabs button{background:none;border:0;padding:10px 17px 12px;color:#829074;min-height:44px;font-size:18px;line-height:1.4}.example-tabs button:first-child{padding-left:0;padding-right:20px}.example-tabs button[aria-selected=true]{text-decoration:underline 2px #4c7141;text-underline-offset:6px;color:#314529}.example-prompt{margin:0 0 18px;font-size:15px;line-height:1.6;overflow-wrap:anywhere}.example-pair{display:grid;grid-template-columns:1fr 1fr;gap:26px}.example-label{display:flex;align-items:center;justify-content:space-between;gap:12px;font-size:12px;margin-bottom:12px}.example-label b{font-size:10px;font-weight:400;margin-left:5px;opacity:.8}.example-model{display:inline-flex;align-items:center;gap:7px;min-width:0;color:#819074}.example-model img{flex:none;object-fit:contain}.example-label>.example-pass,.example-label>.example-fail{flex:none}.example-pass{color:#487347}.example-fail{color:#9b6b57}.example-image{display:block;position:relative;padding:0;background:#e5e8e3;border:0;width:100%;overflow:hidden}.example-image img{display:block;width:100%;height:auto;aspect-ratio:1.35;object-fit:contain;transition:transform .3s}.example-image:hover img{transform:scale(1.04)}.example-image>span{position:absolute;right:14px;bottom:12px;display:flex;align-items:center;gap:7px;font-size:10px;color:#78856f}.example-pair article>p{font-size:12px;margin-top:14px;max-width:450px}@media(max-width:960px){.bench-intro{gap:28px;grid-template-columns:1.15fr 1fr}.bench-intro h1{font-size:44px}.bench-facts{gap:14px;font-size:9px}.score-head,.score-row,.score-axis{grid-template-columns:190px minmax(0,1fr) 85px;gap:18px}.score-number svg{margin-left:6px}.issue-grid{grid-template-columns:repeat(2,minmax(0,1fr))}}@media(max-width:700px){.bench-intro{grid-template-columns:1fr;gap:25px;padding-bottom:30px}.bench-intro h1{font-size:46px}.intro-explanation{max-width:none}.intro-asset-strip{max-width:460px}.intro-assets>span{text-align:left}.bench-facts{font-size:9px;gap:16px}.home-section-heading{margin-bottom:19px;align-items:baseline;flex-wrap:wrap}.home-section-heading h2{font-size:20px}.home-section-heading>span{font-size:10px}.score-head,.score-row,.score-axis{grid-template-columns:145px minmax(0,1fr) 55px;gap:10px}.score-head{font-size:10px}.score-model{gap:8px}.score-model img{width:20px;height:20px}.score-rank{width:11px}.score-model strong{font-size:12px}.score-model small{font-size:9px}.score-number{font-size:18px}.score-number svg{display:none}.score-track{height:7px}.score-axis{font-size:8px}.score-axis span:nth-child(2),.score-axis span:nth-child(4){visibility:hidden}.release-notes{line-height:1.7}.issue-section,.example-section{margin-top:30px;padding-top:28px}.issue-grid{gap:25px 20px}.issue-frequency strong{font-size:25px}.issue-frequency>span{display:block;font-size:9px}.example-pair{gap:12px}.example-image img{aspect-ratio:1}.example-image>span{font-size:9px;right:8px;bottom:8px}.example-image svg{display:none}.example-label{font-size:11px;gap:5px}.example-pair article>p{font-size:11px}.example-tabs button{padding:10px 12px;font-size:16px}.example-prompt{font-size:13px}}@media(prefers-reduced-motion:reduce){.intro-asset-strip img,.example-image img{transition:none}}.asset-preview{position:relative;width:100%;aspect-ratio:1;background:#fff;border:1px solid #e2e2df;overflow:hidden}.asset-canvas{position:absolute;inset:0}.asset-canvas canvas{display:block;width:100%;height:100%;cursor:grab;touch-action:none}.asset-canvas canvas:active{cursor:grabbing}.asset-canvas canvas:focus-visible{outline:2px solid #b9c8ad;outline-offset:-3px}.asset-poster{position:absolute;inset:0;width:100%;height:100%;object-fit:contain;pointer-events:none}.asset-status{position:absolute;top:10px;left:10px;right:10px;color:#777;background:#fff;font-size:11px;padding:4px;pointer-events:none}.asset-hint{position:absolute;top:10px;left:12px;color:#5c685d;font-size:10px;pointer-events:none}@media(max-width:700px){.example-pair{grid-template-columns:1fr;gap:24px}}.example-checks{margin-top:15px;border-top:1px solid #dfe2db}.example-checks-heading{display:flex;justify-content:space-between;align-items:center;padding:12px 0 6px;font-size:11px;color:#73796f}.example-checks-heading button{display:flex;align-items:center;gap:4px;border:0;background:none;padding:3px 0;font:inherit;color:inherit;cursor:pointer}.example-checks-heading button:hover{color:#263329}.example-checks ul{padding:0;margin:0;list-style:none}.example-checks li{display:flex;justify-content:space-between;gap:12px;padding:9px 0;border-bottom:1px solid #e5e7e1;font-size:12px}.check-result{display:inline-flex;align-items:center;gap:7px;font-size:11px}.check-pass{color:#527352}.check-fail{color:#97614f}.check-unavailable{color:#777b73}.example-ocr.example-check-button{align-items:flex-start;border-bottom:1px solid #e5e7e1;font-size:12px}.example-ocr>.check-result{font-size:11px}.check-reading{display:block;margin-top:4px;font-size:12px;line-height:1.4}.example-check-observed.example-check-button{align-items:flex-start}.site-header .brand{font-size:16px;letter-spacing:-.35px}.bench-intro{padding:24px 0 48px}.bench-intro h1{font-size:clamp(48px,6vw,76px);letter-spacing:-.7px;line-height:1;margin-bottom:22px}.intro-description{font-size:24px;letter-spacing:-.6px}.intro-explanation{max-width:430px}.reading-label{color:#77816f;font:11px var(--font-text);margin-bottom:14px;display:block}.example-section{margin-top:0}@media(max-width:700px){.site-header{padding:0 18px;height:70px}.site-header .brand{font-size:12px;gap:7px;letter-spacing:-.3px}.site-header nav{gap:16px}.site-header nav a{font-size:11px}.bench-intro h1{font-size:58px}.intro-description{font-size:21px}}.team-note{max-width:410px;margin:15px 0 0;font-size:12px;line-height:1.7;color:#737c6c}.team-note a,.inline-link{color:#344d36;text-decoration:underline;text-underline-offset:3px}.team-note a:hover,.inline-link:hover{color:#172a19}.bench-intro{grid-template-columns:minmax(0,720px) minmax(0,1fr)}@media(max-width:960px){.bench-intro{grid-template-columns:1fr}.hero-model{display:none}}.hero-model{margin:-40px 0 -28px;justify-self:center;width:min(100%,320px);opacity:0;transition:opacity .6s ease}.hero-model[data-ready]{opacity:1}.hero-model-stage{aspect-ratio:5 / 4;cursor:grab;touch-action:pan-y}.hero-model-stage:active{cursor:grabbing}.hero-model-stage canvas{display:block;width:100%;height:100%}.hero-model figcaption{display:flex;justify-content:center;gap:8px;margin-top:-6px;font:12px var(--font-text);color:#77816f;white-space:nowrap}.hero-model-verdict{padding:0;border:0;background:none;font:inherit;cursor:pointer;text-decoration:none;text-underline-offset:3px}.hero-model-verdict.pass{color:#4c7a4f}.hero-model-verdict.fail{color:#a3483a}.hero-model-verdict:hover:enabled,.hero-model-verdict:focus-visible{text-decoration:underline dotted}.hero-model-verdict:active:enabled{text-decoration:underline solid}.hero-model-verdict:disabled{cursor:default}.hero-model-verdict[aria-busy]{cursor:progress;opacity:.6}@media(prefers-reduced-motion:reduce){.hero-model{transition:none}}.intro-explanation{font-size:15px;max-width:560px;color:#4f5a4c}.team-note{font-size:14px;max-width:560px;color:#5d6759}.bench-facts{font:13px var(--font-text);color:#5d6759;gap:28px}.bench-facts b{font:500 16px var(--font-text);color:#253a2a;margin-right:4px}.board-heading{display:flex;align-items:center;justify-content:space-between;gap:16px;flex-wrap:wrap;margin-bottom:14px}.board-heading h2{font-size:21px;font-weight:500;letter-spacing:-.6px;margin:0}.board-switch{display:inline-flex;border:1px solid #cfd8ca;padding:2px;background:#fff}.board-switch button{border:0;background:none;padding:7px 14px;font-size:13px;color:#5d6759}.board-switch button[aria-pressed=true]{background:#2f4a35;color:#f3f6f1}.board-switch button:focus-visible{outline:2px solid #2f4a35;outline-offset:2px}.board-note{font-size:14px;color:#4f5a4c;max-width:680px;margin:0 0 22px}.score-row-static{cursor:default}.score-row-static:hover{background:none}.score-model small,.score-head{font-size:12px;color:#6b7466}.score-axis>div{font-size:10px;color:#7d8777}.board-details{margin-top:18px;border-top:0;padding-top:0;font-size:13px;color:#4f5a4c}.board-details summary{color:#3d5a43}.board-details table{margin-top:10px;font-size:13px}.board-details td{padding-top:11px;padding-bottom:11px;font-variant-numeric:tabular-nums}.board-fine{font-size:12px;color:#6b7466;margin-top:12px;max-width:680px}.home-section-heading>span,.home-section-heading>a{font-size:13px;color:#5d6759}.issue-frequency>span{font-size:12px;color:#6b7466}.issue-frequency strong{font-family:var(--font-text);font-weight:400;font-size:26px}.example-pair article>p{font-size:13px;color:#5d6759}@media(max-width:700px){.board-heading{align-items:flex-start;flex-direction:column}.board-switch{width:100%}.board-switch button{flex:1}.bench-facts{flex-wrap:wrap;gap:14px 22px;white-space:normal}}.bench-intro{padding-bottom:36px}.bench-intro .intro-description{margin-bottom:0}.about-bench{padding:36px 0 8px;margin-top:28px;border-top:1px solid #dce1da}.about-bench .intro-explanation{margin-top:0}.rank-table{border:1px solid #dde3d9;background:#fff;padding:4px 20px 8px}.rank-title{padding:10px 0 8px;border-bottom:1px solid #dde3d9}.rank-title h3{margin:0;font-size:15px;font-weight:500}.rank-title p{margin:2px 0 0;font-size:12px;color:#6b7466}.rank-row{display:grid;grid-template-columns:minmax(170px,220px) minmax(0,1fr) 118px 78px 132px;align-items:center;gap:20px;width:100%;padding:9px 0;border:0;border-bottom:1px solid #edf0ea;background:none;text-align:left;color:inherit;font:inherit}.rank-row:last-child{border-bottom:0}.rank-head{font-size:12px;color:#6b7466;padding:12px 0 10px;border-bottom-color:#dde3d9}.rank-head span:nth-child(n+3){text-align:right}.rank-link:hover{background:#f4f7f2}.rank-model{display:flex;align-items:center;gap:10px;min-width:0}.rank-model img{object-fit:contain;flex-shrink:0}.rank-model strong{font-size:16px;font-weight:500;white-space:nowrap;overflow:hidden;text-overflow:ellipsis}.effort{color:#6b7466;font-weight:400}.rank-bar{position:relative;height:12px;background:#f0f3ee}.rank-bar i{position:absolute;inset:0 auto 0 0;background:#7f9a78}.rank-head+.rank-row .rank-bar i{background:#3f7a4d}.rank-bar b{position:absolute;top:50%;height:7px;margin-top:-4px;border-left:1px solid #2c3a2e;border-right:1px solid #2c3a2e}.rank-bar b:after{content:"";position:absolute;left:0;right:0;top:3px;border-top:1px solid #2c3a2e}.rank-rate{text-align:right;font:500 16px var(--font-text);color:#1f2a22;font-variant-numeric:tabular-nums;white-space:nowrap}.rank-rate small{display:inline;margin:0;font-size:13px;font-weight:400;color:#7a8475}.rank-extra{text-align:right;font:15px var(--font-text);color:#4f5a4c;font-variant-numeric:tabular-nums}@media(max-width:700px){.rank-table{padding:2px 12px 6px}.rank-row{grid-template-columns:minmax(0,1fr) 72px 96px;gap:12px}.rank-extra{display:none}.rank-model strong{font-size:15px}}.rank-head .rank-extra{font:14px var(--font-text);color:#6b7466;white-space:nowrap}.about-bench .intro-explanation,.about-bench .team-note{max-width:none}.explainer{margin-top:28px;border-top:1px solid #dce1da}.explainer-block{padding:40px 0;border-bottom:1px solid #dce1da}.explainer-block:last-child{border-bottom:0}.explainer-block h2{font-size:24px;font-weight:500;letter-spacing:-.7px;margin:0 0 14px}.explainer-block>p{font-size:15px;line-height:1.75;color:#4f5a4c;max-width:760px}.explainer-block>p+p{margin-top:12px}.explainer-lead{font-size:18px!important;line-height:1.6!important;color:#253028!important;letter-spacing:-.2px}.brief-list{display:grid;grid-template-columns:170px minmax(0,1fr);margin:6px 0 0;border-top:1px solid #dde3d9;max-width:900px}.brief-list dt,.brief-list dd{margin:0;padding:14px 0;border-bottom:1px solid #edf0ea;font-size:14px;line-height:1.65}.brief-list dt{color:#1f2a22;font-weight:500}.brief-list dd{color:#4f5a4c}.explainer-more{display:inline-block;margin-top:18px;font-size:14px}@media(max-width:700px){.explainer-block{padding:30px 0}.explainer-lead{font-size:16px!important}.brief-list{grid-template-columns:1fr}.brief-list dt{padding-bottom:2px;border-bottom:0}.brief-list dd{padding-top:0}}.example-pair{grid-template-columns:repeat(3,minmax(0,1fr));gap:20px}.example-label{font-size:13px}.example-label>.example-pass,.example-label>.example-fail{font-weight:500}.example-pair .asset-hint{display:none}@media(max-width:960px){.example-pair{grid-template-columns:1fr;gap:28px}}.example-models{display:none}@media(max-width:960px){.example-models{display:grid;grid-template-columns:repeat(3,minmax(0,1fr));margin-bottom:14px;border:1px solid #dfe2db}.example-models button{display:flex;align-items:center;justify-content:center;gap:5px;min-width:0;padding:9px 4px;border:0;border-left:1px solid #dfe2db;border-radius:0;background:#fff;font:12px var(--font-text);color:#5d6759;cursor:pointer}.example-models button:first-child{border-left:0}.example-models button>span:nth-child(2){overflow:hidden;text-overflow:ellipsis;white-space:nowrap}.example-models button[aria-selected=true]{background:#f0f3eb;color:#1f2a22;box-shadow:inset 0 -2px #527352}.example-models button:focus-visible{outline:2px solid #527352;outline-offset:-2px}.example-pair article:not([data-current]){display:none}}.metric-section{margin-top:18px}.metric-panels{display:grid;grid-template-columns:repeat(2,minmax(0,1fr));gap:18px}.metric-panels[data-count="1"]{grid-template-columns:minmax(0,1fr)}.metric-panel{border:1px solid #dde3d9;background:#fff;padding:14px 20px 10px;min-width:0}.cost-panel{margin-top:18px}.cost-panel header{display:flex;flex-wrap:wrap;align-items:flex-start;justify-content:space-between;gap:6px 16px}.cost-legend{display:flex;align-items:center;gap:7px}.cost-legend i{width:22px;border-top:1.5px dashed #3f7a4d}.cost-switch{display:flex;gap:2px;margin:-6px -8px 0 0}.cost-switch button{background:none;border:0;padding:6px 8px;min-height:32px;font:13px var(--font-text);color:#6b7466;cursor:pointer;text-underline-offset:5px}.cost-switch button[aria-pressed=true]{color:#314529;text-decoration:underline 1.5px #4c7141}.cost-switch button:hover{color:#314529}.cost-switch button:focus-visible{outline:1px solid #53754e;outline-offset:1px}.cost-frame{position:relative;padding:22px 18px 52px 70px}.cost-axis{position:absolute;font:12px var(--font-text);color:#6b7466;white-space:nowrap}.cost-axis-x{left:70px;right:18px;bottom:6px;text-align:center}.cost-axis-y{left:8px;top:172px;transform:translate(-50%,-50%) rotate(-90deg);transform-origin:center;margin-left:8px}.cost-plot{position:relative;height:300px}.cost-grid{position:absolute;pointer-events:none}.cost-grid-y{left:0;right:0;border-top:1px solid #eef1ec}.cost-grid-y small{position:absolute;right:calc(100% + 8px);top:-8px;font:12px var(--font-text);color:#8a9384;white-space:nowrap}.cost-grid-x{top:0;bottom:0;border-left:1px dashed #eef1ec}.cost-grid-x small{position:absolute;top:calc(100% + 8px);left:0;transform:translate(-50%);font:12px var(--font-text);color:#8a9384;white-space:nowrap}.cost-frontier{position:absolute;inset:0;width:100%;height:100%;overflow:visible}.cost-frontier polyline{fill:none;stroke:#3f7a4d;stroke-width:1.5;stroke-dasharray:4 3}.cost-point{position:absolute;width:16px;height:16px;transform:translate(-50%,-50%)}.cost-anchor{position:absolute;width:0;border-left:1px solid #b9c2b4}.cost-anchor i{position:absolute;left:-3.5px;width:6px;height:6px;margin-top:-3px;border-radius:50%;background:#6b7466}.cost-point img{display:block;width:16px;height:16px;border-radius:3px;background:#fff}.cost-name{position:absolute;font:13px var(--font-text);line-height:15px;color:#6b7466;white-space:nowrap;text-shadow:0 0 3px #fff,0 0 3px #fff,0 0 2px #fff}.cost-point[data-front] .cost-name{color:#222}.cost-point[data-side=right] .cost-name{left:calc(100% + 5px);top:50%;transform:translateY(-50%)}.cost-point[data-side=left] .cost-name{right:calc(100% + 5px);top:50%;transform:translateY(-50%)}.cost-point[data-side=above] .cost-name{bottom:calc(100% + 5px);left:50%;transform:translate(-50%)}.cost-point[data-side=below] .cost-name{top:calc(100% + 5px);left:50%;transform:translate(-50%)}.cost-point:hover{z-index:1}.cost-tip{display:none;position:absolute;left:0;bottom:calc(100% + 8px);width:max-content;max-width:230px;padding:8px 10px;border:1px solid #cbd3c5;background:#fff;color:#263329;font:12px/1.5 var(--font-text);pointer-events:none}.cost-point:hover .cost-tip{display:flex;flex-direction:column}.cost-tip strong{font-weight:500}.cost-point[data-tip-x=end] .cost-tip{left:auto;right:0}.cost-point[data-tip-y=below] .cost-tip{bottom:auto;top:calc(100% + 8px)}@media(max-width:700px){.cost-frame{padding:18px 10px 50px 62px}.cost-axis-x{left:62px;right:10px}.cost-axis-y{top:148px}.cost-plot{height:260px}}.metric-panel header{padding-bottom:8px;border-bottom:1px solid #dde3d9}.metric-panel h3{margin:0;font-size:15px;font-weight:500}.metric-panel header p{margin:2px 0 0;font-size:12px;color:#6b7466}.elo-panel header{display:flex;align-items:flex-start;justify-content:space-between;gap:12px}.elo-cycle{display:inline-flex;flex-shrink:0;align-items:center;margin:-6px -8px 0 0;background:none;border:0;border-radius:0;padding:6px 8px;min-height:32px;font-size:13px;color:#314529;cursor:pointer;text-decoration:underline 1.5px #4c7141;text-underline-offset:5px}.elo-cycle:focus-visible{outline:1px solid #53754e;outline-offset:1px}.metric-row,.metric-axis{display:grid;grid-template-columns:minmax(110px,150px) minmax(0,1fr) 68px;align-items:center;gap:14px}.metric-row{height:30px;border-bottom:1px solid #edf0ea}.metric-row strong{font-size:13px}.metric-value{text-align:right;font:500 13px var(--font-text);color:#1f2a22;font-variant-numeric:tabular-nums}.score-tooltip{position:relative}.metric-hit{stroke:transparent;stroke-width:12;pointer-events:stroke;vector-effect:non-scaling-stroke}.score-tooltip:focus-visible{outline:1px solid #53754e;outline-offset:3px}.score-tooltip-popup{position:absolute;z-index:100;right:0;bottom:calc(100% + 8px);width:max-content;max-width:min(290px,calc(100vw - 40px));padding:8px 10px;border:1px solid #cbd3c5;background:#fff;color:#263329;text-align:left;font:12px/1.5 var(--font-text);white-space:normal}.score-tooltip-popup:not([hidden]){display:flex;flex-direction:column}.score-tooltip-popup strong{font-size:inherit;font-weight:500}.score-tooltip-popup:after{content:"";position:absolute;left:0;right:0;bottom:-12px;height:12px}.score-tooltip-popup[data-placement=below]{top:calc(100% + 8px);bottom:auto}.score-tooltip-popup[data-placement=below]:after{top:-12px;bottom:auto}.metric-axis{padding-top:4px}.metric-axis>span:nth-child(2){position:relative;height:14px;font:10px var(--font-text);color:#96a08b}.metric-axis small{position:absolute;top:0;transform:translate(-50%);font-size:inherit}@media(max-width:900px){.metric-panels{grid-template-columns:minmax(0,1fr)}}@media(max-width:700px){.metric-panel{padding:12px 12px 8px}.metric-row,.metric-axis{grid-template-columns:minmax(0,1fr) 96px 68px;gap:10px}.metric-axis small:not(:first-child):not(:last-child){display:none}}.metric-dots{display:block;height:18px}.metric-dots svg{display:block;overflow:visible}.metric-dots circle{fill:#3f7a4d;fill-opacity:.16}.metric-zero{stroke:#b9c2b4;stroke-width:1;stroke-dasharray:2 2}.metric-mean{stroke:#1f2a22;stroke-width:2.5}.metric-interval{display:none}@media(hover:none){.metric-interval{display:block;margin:0;font-size:11px;font-weight:400;line-height:1;color:#7a8475}}.example-checks li{padding:0}.example-check-button{position:relative;display:flex;justify-content:space-between;align-items:center;gap:12px;width:100%;border:0;background:transparent;padding:9px 6px;text-align:left;font:inherit;color:inherit;cursor:pointer;border-radius:0}.example-check-button:hover{background:#f0f3eb}.example-check-button[aria-pressed=true]{background:#e6eddf}.example-check-button[aria-pressed=true]:before{content:"";position:absolute;left:0;top:-1px;bottom:-1px;width:3px;background:#527352;pointer-events:none}.example-check-button:focus-visible{outline:2px solid #527352;outline-offset:2px}.check-isolation-hint{margin:0 0 6px;font-size:11px;color:#73796f}.asset-focus-status{position:absolute;top:10px;left:10px;right:10px;width:fit-content;max-width:calc(100% - 20px);padding:5px 8px;border-radius:0;background:#fbfcf9e8;color:#43593c;font-size:11px;pointer-events:none}.metric-head{color:#697566;font-size:11px}.metric-head>:last-child{text-align:right}.citation-code{max-width:760px;margin:18px 0 0;white-space:pre-wrap;overflow-wrap:anywhere}.model-evaluation-contact{margin:18px 0 0;font-size:14px;line-height:1.6;color:#4f5a4c}.model-evaluation-contact a{white-space:nowrap}.elo-panel{container-type:inline-size}.elo-bars{--plot: 250px;display:grid;grid-auto-flow:column;grid-auto-columns:minmax(0,1fr);column-gap:6px;margin-top:30px}.elo-column{position:relative;display:flex;flex-direction:column;min-width:0;cursor:pointer}.elo-plot{order:-1;position:relative;height:var(--plot)}.elo-bar{position:absolute;bottom:0;left:0;width:100%;height:calc(var(--value) * 100%);background:color-mix(in srgb,var(--bar) 24%,#fff);border-top:2px solid var(--bar)}.elo-logo{position:absolute;bottom:5px;left:0;right:0;display:grid;place-items:center;width:18px;height:18px;margin:0 auto;border-radius:4px;background:#fff}.elo-logo img{width:13px;height:13px;object-fit:contain;filter:brightness(1.04);mix-blend-mode:multiply}.elo-score{position:absolute;left:0;right:0;top:calc((1 - var(--value)) * var(--plot) - 5px);transform:translateY(-100%);text-align:center;font:500 13px/1.2 var(--font-text);color:#1f2a22;font-variant-numeric:lining-nums tabular-nums;white-space:nowrap}.elo-error{position:absolute;top:calc(100% + 11px);left:0;right:0;font-size:10px;font-weight:400;color:#4f5a4c;opacity:0;transition:opacity .15s;pointer-events:none}.elo-column:is(:hover,:focus) .elo-error{opacity:1}.elo-bar{transition:background-color .15s}.elo-column:is(:hover,:focus) .elo-bar{background:color-mix(in srgb,var(--bar) 38%,#fff)}.elo-column:focus{outline:none}.elo-column:focus-visible{outline:1px solid #53754e;outline-offset:3px}.elo-model{display:flex;flex-direction:column;align-items:center;padding-top:9px;text-align:center}.elo-model img{display:none;flex-shrink:0;object-fit:contain}.elo-model strong{font:500 11px/1.2 var(--font-text);color:#263329}@container (max-width: 380px){.elo-bars{grid-auto-flow:row;margin-top:4px}.elo-column{flex-direction:row;align-items:center;gap:10px;min-height:38px;border-bottom:1px solid #edf0ea}.elo-model{flex-direction:row;gap:8px;width:112px;flex-shrink:0;padding:0;text-align:left}.elo-model img{display:block}.elo-model strong{font-size:13px}.elo-plot{order:0;flex:1;height:12px}.elo-bar{top:0;bottom:auto;width:calc(var(--value) * 100%);height:100%;border-top:0;border-right:2px solid var(--bar)}.elo-logo{display:none}.elo-score{position:relative;top:auto;left:auto;right:auto;width:40px;flex-shrink:0;transform:none;text-align:right}.elo-error{top:calc(100% - 2px);left:auto;right:0;font-size:11px;line-height:1}}
