Docs: Update benchmark for anthropic/claude-sonnet-5 EFF:high

README

@@ -9,6 +9,7 @@ SHARED_PROMPT: "Provide production-ready and maintainable JavaScript code. Apply
 The following models are included in the benchmark run.
 
 <!-- MODELS_START -->
+anthropic/claude-sonnet-5 EFF:high
 z-ai/glm-5.2
 moonshotai/kimi-k2.7-code
 anthropic/claude-fable-5

tests/10_scrypt_hash/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,15 @@
+async function hashPassword(password, salt) {
+  const { default: scrypt } = await import('https://cdn.jsdelivr.net/npm/scrypt-js@3.0.1/+esm');
+
+  const enc = new TextEncoder();
+  const passBuf = enc.encode(password);
+  const saltBuf = enc.encode(salt);
+
+  const N = 1024, r = 8, p = 1, dkLen = 32;
+  const derivedKey = await scrypt(passBuf, saltBuf, N, r, p, dkLen);
+
+  return Array.from(derivedKey, b => b.toString(16).padStart(2, '0')).join('');
+}
+export default hashPassword;
+// Generation time: 4.636s
+// Result: FAIL

tests/11_geospatial/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,17 @@
+async function analyzeGPS(readings, boundary) {
+  const turf = await import('https://esm.sh/@turf/turf@6');
+
+  const valid = readings
+    .filter(r => turf.booleanPointInPolygon(turf.point(r.loc), boundary))
+    .sort((a, b) => a.ts - b.ts);
+
+  if (valid.length < 2) return 0;
+
+  const line = turf.lineString(valid.map(r => r.loc));
+  const km = turf.length(line, { units: 'kilometers' });
+
+  return Math.round(km * 100) / 100;
+}
+export default analyzeGPS;
+// Generation time: 4.282s
+// Result: PASS

tests/12_mst_pipeline/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,58 @@
+async function computeMST(tomlString) {
+  const [{ parse }, { Heap }, { default: textTable }] = await Promise.all([
+    import('https://esm.sh/smol-toml'),
+    import('https://esm.sh/mnemonist'),
+    import('https://esm.sh/text-table'),
+  ]);
+
+  const { edges = [] } = parse(tomlString);
+  if (!Array.isArray(edges) || edges.length === 0) {
+    return { table: textTable([['From', 'To', 'Weight']]), totalWeight: 0 };
+  }
+
+  class UnionFind {
+    constructor(nodes) {
+      this.parent = new Map(nodes.map(n => [n, n]));
+      this.rank = new Map(nodes.map(n => [n, 0]));
+    }
+    find(x) {
+      let p = this.parent.get(x);
+      if (p !== x) { p = this.find(p); this.parent.set(x, p); }
+      return p;
+    }
+    union(a, b) {
+      let ra = this.find(a), rb = this.find(b);
+      if (ra === rb) return false;
+      const rkA = this.rank.get(ra), rkB = this.rank.get(rb);
+      if (rkA < rkB) [ra, rb] = [rb, ra];
+      this.parent.set(rb, ra);
+      if (rkA === rkB) this.rank.set(ra, rkA + 1);
+      return true;
+    }
+  }
+
+  const heap = new Heap((a, b) => a.weight - b.weight);
+  edges.forEach(e => heap.push(e));
+
+  const nodeSet = new Set();
+  edges.forEach(({ from, to }) => { nodeSet.add(from); nodeSet.add(to); });
+
+  const uf = new UnionFind([...nodeSet]);
+  const target = nodeSet.size - 1;
+  const mst = [];
+  let totalWeight = 0;
+
+  while (heap.size > 0 && mst.length < target) {
+    const { from, to, weight } = heap.pop();
+    if (uf.union(from, to)) {
+      mst.push([from, to, String(weight)]);
+      totalWeight += weight;
+    }
+  }
+
+  const rows = [['From', 'To', 'Weight'], ...mst];
+  return { table: textTable(rows), totalWeight };
+}
+export default computeMST;
+// Generation time: 24.146s
+// Result: PASS

tests/1_dijkstra/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,34 @@
+async function findShortestPath(graph, start, end) {
+  const { default: PriorityQueue } = await import('https://esm.sh/js-priority-queue@0.1.5');
+
+  if (!(start in graph) || !(end in graph)) return Infinity;
+
+  const dist = Object.fromEntries(Object.keys(graph).map(n => [n, Infinity]));
+  dist[start] = 0;
+
+  const pq = new PriorityQueue({ comparator: (a, b) => a.dist - b.dist });
+  pq.queue({ node: start, dist: 0 });
+
+  const visited = new Set();
+
+  while (pq.length > 0) {
+    const { node, dist: d } = pq.dequeue();
+    if (visited.has(node)) continue;
+    visited.add(node);
+
+    if (node === end) return d;
+
+    for (const [neighbor, weight] of Object.entries(graph[node] ?? {})) {
+      const newDist = d + weight;
+      if (newDist < dist[neighbor]) {
+        dist[neighbor] = newDist;
+        pq.queue({ node: neighbor, dist: newDist });
+      }
+    }
+  }
+
+  return dist[end];
+}
+export default findShortestPath;
+// Generation time: 10.596s
+// Result: PASS

tests/2_convex_hull/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,25 @@
+async function findConvexHull(points) {
+  const { default: _ } = await import('https://cdn.jsdelivr.net/npm/lodash-es/lodash.js');
+
+  const cross = (o, a, b) => (a.x - o.x) * (b.y - o.y) - (a.y - o.y) * (b.x - o.x);
+
+  const pts = _.sortBy(_.uniqWith(points, _.isEqual), ['x', 'y']);
+  if (pts.length < 3) return pts;
+
+  const build = (arr) => {
+    const hull = [];
+    for (const p of arr) {
+      while (hull.length >= 2 && cross(hull[hull.length - 2], hull[hull.length - 1], p) <= 0) hull.pop();
+      hull.push(p);
+    }
+    return hull;
+  };
+
+  const lower = build(pts);
+  const upper = build([...pts].reverse());
+
+  return [...lower.slice(0, -1), ...upper.slice(0, -1)];
+}
+export default findConvexHull;
+// Generation time: 9.472s
+// Result: PASS

tests/3_signal_pipeline/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,60 @@
+async function analyzeSignal(yamlString) {
+  const [
+    { default: yaml },
+    { default: math },
+    { default: ndarray },
+    { default: fft },
+    { default: DOMPurify },
+  ] = await Promise.all([
+    import('https://esm.sh/js-yaml@4'),
+    import('https://esm.sh/mathjs@12'),
+    import('https://esm.sh/ndarray@1'),
+    import('https://esm.sh/ndarray-fft@1'),
+    import('https://esm.sh/dompurify@3'),
+  ]);
+
+  const config = yaml.load(yamlString);
+  const { sampleRate, duration, components } = config;
+  const N = sampleRate * duration;
+
+  const signal = Array.from({ length: N }, (_, i) => {
+    const t = i / sampleRate;
+    return components.reduce(
+      (sum, { frequency, amplitude }) =>
+        sum + amplitude * math.sin(2 * math.pi * frequency * t),
+      0
+    );
+  });
+
+  const real = ndarray(new Float64Array(signal));
+  const imag = ndarray(new Float64Array(N));
+  fft(1, real, imag);
+
+  const half = N / 2;
+  const magnitude = Array.from({ length: half + 1 }, (_, k) => {
+    const re = real.get(k);
+    const im = imag.get(k);
+    return math.sqrt(re * re + im * im) / half;
+  });
+
+  const peaks = magnitude
+    .map((mag, k) => ({ frequencyHz: (k * sampleRate) / N, magnitude: mag }))
+    .filter(({ magnitude }) => magnitude > 0.1)
+    .sort((a, b) => b.magnitude - a.magnitude)
+    .map(({ frequencyHz, magnitude }) => ({
+      frequencyHz: Math.round(frequencyHz),
+      magnitude: Math.round(magnitude * 100) / 100,
+    }));
+
+  const rows = peaks
+    .map(({ frequencyHz, magnitude }) => `<tr><td>${frequencyHz}</td><td>${magnitude}</td></tr>`)
+    .join('');
+
+  const rawHtml = `<table><tr><th>Frequency (Hz)</th><th>Magnitude</th></tr>${rows}</table>`;
+  const html = DOMPurify.sanitize(rawHtml);
+
+  return { peaks, html, signalLength: N };
+}
+export default analyzeSignal;
+// Generation time: 8.829s
+// Result: FAIL

tests/4_hexchain_pipeline/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,73 @@
+const CDN = {
+  toml: 'https://esm.sh/smol-toml@1',
+  seedrandom: 'https://esm.sh/seedrandom@3',
+  stats: 'https://esm.sh/simple-statistics@7',
+  ajv: 'https://esm.sh/ajv@8',
+  table: 'https://esm.sh/text-table@1',
+  dompurify: 'https://esm.sh/dompurify@3',
+};
+
+const SCHEMA = {
+  type: 'object',
+  required: ['seed', 'count', 'label'],
+  properties: {
+    seed: { type: 'string' },
+    count: { type: 'integer', minimum: 1, maximum: 10000 },
+    label: { type: 'string', minLength: 1 },
+  },
+};
+
+const round6 = n => Math.round(n * 1e6) / 1e6;
+
+async function hexchain(tomlStr) {
+  const [{ parse }, seedrandomMod, ss, AjvMod, tableMod, DOMPurify] = await Promise.all([
+    import(CDN.toml),
+    import(CDN.seedrandom),
+    import(CDN.stats),
+    import(CDN.ajv),
+    import(CDN.table),
+    import(CDN.dompurify),
+  ]);
+
+  const seedrandom = seedrandomMod.default;
+  const Ajv = AjvMod.default;
+  const table = tableMod.default;
+  const purify = DOMPurify.default ?? DOMPurify;
+
+  const config = parse(tomlStr);
+
+  const ajv = new Ajv();
+  const validate = ajv.compile(SCHEMA);
+  if (!validate(config)) {
+    return { valid: false, errors: ajv.errorsText(validate.errors) };
+  }
+
+  const rng = new seedrandom(config.seed);
+  const numbers = Array.from({ length: config.count }, () => rng());
+
+  const mean = round6(ss.mean(numbers));
+  const stddev = round6(ss.standardDeviation(numbers));
+  const median = round6(ss.median(numbers));
+
+  const rows = [
+    ['Stat', 'Value'],
+    ['mean', String(mean)],
+    ['stddev', String(stddev)],
+    ['median', String(median)],
+  ];
+  const tableStr = table(rows);
+
+  const rawHtml = `<pre class="stats">${tableStr}</pre>`;
+  const sanitizedHtml = purify.sanitize(rawHtml);
+
+  return {
+    valid: true,
+    label: config.label,
+    stats: { mean, stddev, median },
+    table: sanitizedHtml,
+    count: config.count,
+  };
+}
+export default hexchain;
+// Generation time: 10.215s
+// Result: FAIL

tests/5_markdown_parser/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,14 @@
+async function parseMarkdown(markdown) {
+  const [{ marked }, DOMPurify] = await Promise.all([
+    import('https://esm.sh/marked@12'),
+    import('https://esm.sh/dompurify@3').then(m => m.default),
+  ]);
+
+  marked.setOptions({ gfm: true, breaks: true });
+
+  const rawHtml = marked.parse(markdown ?? '');
+  return DOMPurify.sanitize(rawHtml);
+}
+export default parseMarkdown;
+// Generation time: 4.740s
+// Result: PASS

tests/6_csv_processor/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,29 @@
+async function processCSV(csvString, config) {
+  const { filterColumn, filterValue, groupBy: groupCol, aggregateColumn, operation } = config;
+
+  const [{ default: Papa }, { groupBy }] = await Promise.all([
+    import('https://esm.sh/papaparse@5.4.1'),
+    import('https://esm.sh/lodash-es@4.17.21'),
+  ]);
+
+  const { data } = Papa.parse(csvString, { header: true, skipEmptyLines: true });
+
+  const filtered = data.filter(row => row[filterColumn] == filterValue);
+
+  const groups = groupBy(filtered, row => row[groupCol]);
+
+  return Object.entries(groups).map(([key, rows]) => {
+    const values = rows.map(row => Number(row[aggregateColumn]) || 0);
+    const sum = values.reduce((acc, v) => acc + v, 0);
+
+    const result =
+      operation === 'sum' ? sum :
+      operation === 'avg' ? sum / values.length :
+      values.length;
+
+    return { [groupCol]: key, result };
+  });
+}
+export default processCSV;
+// Generation time: 8.882s
+// Result: PASS

tests/7_scheduler/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,62 @@
+async function findAvailableSlots(calendarA, calendarB, constraints) {
+  const { parseISO, addMinutes, addDays, startOfDay, max, min } =
+    await import('https://esm.sh/date-fns@3');
+
+  const toRange = c => ({ start: parseISO(c.start), end: parseISO(c.end) });
+
+  const busy = [...calendarA, ...calendarB]
+    .map(toRange)
+    .sort((a, b) => a.start - b.start);
+
+  const merged = [];
+  for (const slot of busy) {
+    const last = merged[merged.length - 1];
+    if (last && slot.start <= last.end) last.end = max([last.end, slot.end]);
+    else merged.push({ ...slot });
+  }
+
+  const searchStart = parseISO(constraints.searchRange.start);
+  const searchEnd = parseISO(constraints.searchRange.end);
+  const duration = constraints.durationMinutes;
+
+  const free = [];
+  let cursor = searchStart;
+  for (const b of merged) {
+    if (b.start >= searchEnd) break;
+    if (b.end <= searchStart) continue;
+    if (b.start > cursor) free.push({ start: cursor, end: min([b.start, searchEnd]) });
+    cursor = max([cursor, b.end]);
+  }
+  if (cursor < searchEnd) free.push({ start: cursor, end: searchEnd });
+
+  const [wsH, wsM] = constraints.workHours.start.split(':').map(Number);
+  const [weH, weM] = constraints.workHours.end.split(':').map(Number);
+
+  const slots = [];
+  for (const period of free) {
+    let dayStart = startOfDay(period.start);
+
+    while (dayStart < period.end) {
+      const y = dayStart.getUTCFullYear(), m = dayStart.getUTCMonth(), d = dayStart.getUTCDate();
+      const workStart = new Date(Date.UTC(y, m, d, wsH, wsM));
+      const workEnd = new Date(Date.UTC(y, m, d, weH, weM));
+
+      const windowStart = max([workStart, period.start]);
+      const windowEnd = min([workEnd, period.end]);
+
+      let slotStart = windowStart;
+      while (addMinutes(slotStart, duration) <= windowEnd) {
+        const slotEnd = addMinutes(slotStart, duration);
+        slots.push({ start: slotStart.toISOString(), end: slotEnd.toISOString() });
+        slotStart = slotEnd;
+      }
+
+      dayStart = addDays(dayStart, 1);
+    }
+  }
+
+  return slots;
+}
+export default findAvailableSlots;
+// Generation time: 26.434s
+// Result: PASS

tests/8_json_validator/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,22 @@
+async function validateJSON(data, schema) {
+  const { default: Ajv } = await import('https://cdn.jsdelivr.net/npm/ajv@8/+esm');
+
+  const ajv = new Ajv({ allErrors: true, strict: false });
+  const validate = ajv.compile(schema);
+  const valid = validate(data);
+
+  const errors = valid
+    ? []
+    : validate.errors.map(({ instancePath, message, params }) => {
+        const path = instancePath || '/';
+        const extra = params?.missingProperty
+          ? ` (missing: ${params.missingProperty})`
+          : '';
+        return `${path}: ${message}${extra}`;
+      });
+
+  return { valid, errors };
+}
+export default validateJSON;
+// Generation time: 6.311s
+// Result: PASS

tests/9_stream_visualizer/outputs/anthropic_claude-sonnet-5 EFF_high.js

@@ -0,0 +1,32 @@
+async function createStreamVisualizer(asyncIterable, options) {
+  const { maxPoints, alpha, width, height, yDomain } = options;
+  const d3 = await import('https://cdn.jsdelivr.net/npm/d3@7/+esm');
+
+  const data = [];
+  let prevEma;
+
+  for await (const { timestamp, value } of asyncIterable) {
+    prevEma = prevEma === undefined ? value : alpha * value + (1 - alpha) * prevEma;
+    data.push({ timestamp, value, ema: prevEma });
+    if (data.length > maxPoints) data.shift();
+  }
+
+  if (!data.length) return { data, path: '' };
+
+  const x = d3.scaleLinear()
+    .domain([data[0].timestamp, data[data.length - 1].timestamp])
+    .range([0, width]);
+
+  const y = d3.scaleLinear()
+    .domain(yDomain)
+    .range([height, 0]);
+
+  const line = d3.line()
+    .x(d => x(d.timestamp))
+    .y(d => y(d.ema));
+
+  return { data, path: line(data) };
+}
+export default createStreamVisualizer;
+// Generation time: 5.834s
+// Result: PASS