Expansion page: policybench.org/expand

app/src/App.tsx

@@ -386,6 +386,12 @@ export default function App() {
             PolicyBench.org
           </a>{" "}
           &middot;{" "}
+          <a
+            href="/expand"
+            className="text-text-secondary hover:text-primary transition-colors"
+          >
+            Expand
+          </a>
           <a
             href="https://github.com/PolicyEngine/policybench"
             className="text-text-secondary hover:text-primary transition-colors"

app/src/app/expand/page.tsx

@@ -0,0 +1,168 @@
+import type { Metadata } from "next";
+import Link from "next/link";
+
+import SiteHeader from "../../components/SiteHeader";
+
+const DESCRIPTION =
+  "Expand PolicyBench to your region, population, or program: households drawn to your area, references computed from the law, every miss diagnosed, results public.";
+
+export const metadata: Metadata = {
+  title: "Expand",
+  description: DESCRIPTION,
+  alternates: {
+    canonical: "/expand",
+  },
+  openGraph: {
+    type: "website",
+    url: "https://policybench.org/expand",
+    siteName: "PolicyBench",
+    title: "Expand PolicyBench",
+    description: DESCRIPTION,
+    images: [
+      {
+        url: "/og-image.png",
+        width: 1200,
+        height: 630,
+        alt: "PolicyBench — an LLM benchmark for tax and benefit calculation",
+      },
+    ],
+  },
+};
+
+function Package({
+  title,
+  price,
+  children,
+}: {
+  title: string;
+  price: string;
+  children: React.ReactNode;
+}) {
+  return (
+    <div className="card px-5 py-5 flex flex-col">
+      <div className="text-[10px] uppercase tracking-[0.14em] text-text-muted font-medium">
+        {title}
+      </div>
+      <div className="mt-2 font-[family-name:var(--font-display)] text-2xl text-text">
+        {price}
+      </div>
+      <div className="mt-3 text-sm leading-relaxed text-text-secondary">
+        {children}
+      </div>
+    </div>
+  );
+}
+
+export default function ExpandPage() {
+  return (
+    <div className="min-h-screen">
+      <SiteHeader
+        actionLink={{
+          label: "Benchmark",
+          href: "/",
+          type: "internal",
+        }}
+      />
+      <main className="max-w-4xl mx-auto px-4 sm:px-6 pt-12 pb-20">
+        <div className="eyebrow mb-3">PolicyBench &middot; expansion</div>
+        <h1 className="font-[family-name:var(--font-display)] text-4xl md:text-5xl text-text tracking-tight">
+          Expand PolicyBench to your region, population, or program
+        </h1>
+
+        <p className="mt-6 text-base leading-relaxed text-text-secondary">
+          Families already ask AI about the questions that decide their month:
+          Do I qualify for SNAP? How much is my credit? Will this job cost me
+          Medicaid? The{" "}
+          <Link href="/" className="text-primary hover:underline">
+            public board
+          </Link>{" "}
+          tests 30 frontier models on 100 real households.
+          The best model computes 88.7% of amounts within $1. On SNAP cases
+          where the family is owed benefits, models answer exactly $0 in 42% of
+          answers, and no model gets more than 1 case in 20 right.
+          On Medicaid eligibility, the median model misclassifies 1 person in
+          15; the weakest, nearly 1 in 3. A family
+          told &ldquo;$0&rdquo; does not apply. Those are national numbers —
+          nobody measures this for your region.
+        </p>
+
+        <p className="mt-4 text-base leading-relaxed text-text-secondary">
+          A PolicyBench slice measures it. We draw households from
+          survey microdata weighted to your population, cover your programs,
+          benchmark the models behind the tools your people use, and compute every
+          reference from the law with PolicyEngine. Every miss gets a
+          diagnosed failure mode. You get a public slice leaderboard, a
+          written analysis of where models fail your population, and a
+          briefing. The answer key checks itself in public: references come
+          from open-source code, cross-checked against other calculators
+          where they exist, and challenged values get adjudicated against
+          the statute — the{" "}
+          <Link href="/paper" className="text-primary hover:underline">
+            methodology and adjudication record
+          </Link>{" "}
+          are published.
+        </p>
+
+        <div className="grid sm:grid-cols-3 gap-4 mt-10">
+          <Package title="Program deep-dive" price="from $7,500">
+            One program family — SNAP, Medicaid, child care, tax credits —
+            across all 30 board models. Per-model accuracy, diagnosed failure
+            modes, written analysis, and a briefing. Fast: the board already
+            holds the raw material.
+          </Package>
+          <Package title="State or city slice" price="from $20,000">
+            New households weighted to your area and program mix. A published
+            slice leaderboard beside the national board, the full audit, the
+            analysis, and a briefing for your team or grantees.
+          </Package>
+          <Package title="National or portfolio" price="let&rsquo;s talk">
+            A 50-state observatory, or coverage across a whole grantee
+            portfolio — including your grantees&rsquo; own tools, run through
+            the same households as the board. We scope these directly.
+          </Package>
+        </div>
+
+        <p className="mt-8 text-sm leading-relaxed text-text-secondary">
+          We size samples to the confidence width your question needs, from
+          your population — not a fixed household count. Any slice can add
+          standing coverage: quarterly refresh and re-analysis, new models as
+          they ship, and technical assistance to grantees building AI tools —
+          priced with the slice.
+        </p>
+
+        <div className="card px-5 py-4 mt-8">
+          <div className="text-[10px] uppercase tracking-[0.14em] text-text-muted font-medium">
+            What funding buys — and what it never buys
+          </div>
+          <p className="mt-2 text-sm leading-relaxed text-text-secondary">
+            Funders buy coverage: households, programs, regions, refresh
+            cadence. Funding never buys scores, rankings, or placement. No
+            model vendor pays for evaluation — vendor funding for separate
+            projects never touches the benchmark. Every slice stays public —
+            prompts, references, predictions, and diagnoses.
+          </p>
+        </div>
+
+        <div className="mt-10 flex flex-col sm:flex-row sm:items-center gap-4">
+          <a
+            href="mailto:contact@policybench.org?subject=Expanding%20PolicyBench"
+            className="inline-flex items-center justify-center rounded-full bg-primary px-6 py-3 text-sm font-semibold hover:opacity-90 transition-opacity"
+            style={{ color: "var(--background)" }}
+          >
+            Contact us
+          </a>
+          <span className="text-sm text-text-secondary">
+            contact@policybench.org
+          </span>
+        </div>
+        <p className="mt-4 text-base leading-relaxed text-text-secondary">
+          PolicyBench is a project of PolicyEngine, a 501(c)(3) nonprofit —
+          engagements work as grants or contracts. We also brief funder
+          networks — one slice presented to a convened
+          room goes further than a dozen pitches, and we are glad to present
+          at yours.
+        </p>
+      </main>
+    </div>
+  );
+}

app/src/app/sitemap.ts

@@ -21,6 +21,11 @@ export default function sitemap(): MetadataRoute.Sitemap {
       changeFrequency: "monthly",
       priority: 0.7,
     },
+    {
+      url: "https://policybench.org/expand",
+      changeFrequency: "monthly",
+      priority: 0.5,
+    },
     ...modelEntries,
   ];
 }