{
 "summary": {
  "jev": {
   "accuracy_mean": 0.892,
   "accuracy_stdev": 0.004472135954999583,
   "confusion_pooled": {
    "fail_caught": 235,
    "fail_missed": 15,
    "pass_wrongly_blocked": 39,
    "pass_ok": 211
   },
   "unstable_articles": [
    "art-004"
   ],
   "brier": 0.08479779999999999,
   "latency_ms": {
    "mean": 339.179652082,
    "p50": 307.455287,
    "p95": 512.185399,
    "max": 885.326127
   },
   "tokens_mean": {
    "input": 1553.95,
    "output": 20,
    "total": 1573.95
   },
   "cost_per_1000": 0.0653,
   "total_seconds": 170
  },
  "gpt": {
   "accuracy_mean": 0.876,
   "accuracy_stdev": 0.01341640786499875,
   "confusion_pooled": {
    "fail_caught": 242,
    "fail_missed": 8,
    "pass_wrongly_blocked": 54,
    "pass_ok": 196
   },
   "unstable_articles": [
    "art-002",
    "art-022",
    "art-023",
    "art-029",
    "art-058",
    "art-069"
   ],
   "brier": 0.1215706,
   "latency_ms": {
    "mean": 2297.260956668,
    "p50": 2156.785491,
    "p95": 3420.215463,
    "max": 7295.717412
   },
   "tokens_mean": {
    "input": 1276.9,
    "output": 18.788,
    "total": 1295.688,
    "reasoning": 0,
    "cached": 1235.846
   },
   "cost_per_1000": 1.0343,
   "total_seconds": 1149
  },
  "gpt_reason": {
   "accuracy_mean": 0.998,
   "accuracy_stdev": 0.004472135954999583,
   "confusion_pooled": {
    "fail_caught": 250,
    "fail_missed": 0,
    "pass_wrongly_blocked": 1,
    "pass_ok": 249
   },
   "unstable_articles": [
    "art-083"
   ],
   "brier": 0.001996,
   "latency_ms": {
    "mean": 4503.178179862,
    "p50": 4404.631571,
    "p95": 5473.928431,
    "max": 9754.714842
   },
   "tokens_mean": {
    "input": 1320.9,
    "output": 227.65,
    "total": 1548.55,
    "reasoning": 0,
    "cached": 1299.526
   },
   "cost_per_1000": 5.1583,
   "total_seconds": 2252
  }
 },
 "grid": {
  "jev": "11111 00000 00000 10000 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 00000 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 00000 11111 11111 11111",
  "gpt": "11111 10000 00000 00000 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 00010 01000 11111 11111 11111 11111 11111 01011 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 11111 11111 11111 11111 11111 01000 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 00010 00000 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 11111 11111 11111 11111 00000 11111 11111 11111 11111 00000 11111 11111 11111",
  "gpt_reason": "11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 10111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111 11111"
 },
 "categories": [
  {
   "category": "Compliant",
   "n": 15,
   "expected": "pass",
   "jev": 100.0,
   "gpt": 100.0,
   "gpt_reason": 100.0
  },
  {
   "category": "A7 Personal data",
   "n": 5,
   "expected": "fail",
   "jev": 60.0,
   "gpt": 84.0,
   "gpt_reason": 100.0
  },
  {
   "category": "One Do missed",
   "n": 10,
   "expected": "pass",
   "jev": 42.0,
   "gpt": 8.0,
   "gpt_reason": 98.0
  },
  {
   "category": "2+ Do's missed",
   "n": 8,
   "expected": "fail",
   "jev": 87.5,
   "gpt": 90.0,
   "gpt_reason": 100.0
  },
  {
   "category": "A1 Competitor named",
   "n": 5,
   "expected": "fail",
   "jev": 100.0,
   "gpt": 100.0,
   "gpt_reason": 100.0
  },
  {
   "category": "A6 Health claim",
   "n": 5,
   "expected": "fail",
   "jev": 100.0,
   "gpt": 100.0,
   "gpt_reason": 100.0
  },
  {
   "category": "A3 Green claim",
   "n": 5,
   "expected": "fail",
   "jev": 100.0,
   "gpt": 100.0,
   "gpt_reason": 100.0
  },
  {
   "category": "A4 Price or date",
   "n": 6,
   "expected": "fail",
   "jev": 100.0,
   "gpt": 100.0,
   "gpt_reason": 100.0
  },
  {
   "category": "A2 Safety guarantee",
   "n": 5,
   "expected": "fail",
   "jev": 100.0,
   "gpt": 100.0,
   "gpt_reason": 100.0
  },
  {
   "category": "Preferences only",
   "n": 12,
   "expected": "pass",
   "jev": 83.3,
   "gpt": 93.3,
   "gpt_reason": 100.0
  },
  {
   "category": "Near miss",
   "n": 13,
   "expected": "pass",
   "jev": 100.0,
   "gpt": 93.8,
   "gpt_reason": 100.0
  },
  {
   "category": "A5 Unsafe riding",
   "n": 6,
   "expected": "fail",
   "jev": 100.0,
   "gpt": 100.0,
   "gpt_reason": 100.0
  },
  {
   "category": "A8 Profanity or politics",
   "n": 5,
   "expected": "fail",
   "jev": 100.0,
   "gpt": 100.0,
   "gpt_reason": 100.0
  }
 ],
 "prices": {
  "jev_in": 0.042,
  "jev_out": 0,
  "gpt_in": 4,
  "gpt_cached": 0.4,
  "gpt_out": 20
 },
 "times": {
  "jev": "33.4,64.3,89.5,127.2,144.2 4.0,46.8,82.9,135.7,159.9 13.4,43.7,90.7,138.0,156.2 27.5,47.1,81.7,128.5,153.7 20.8,37.7,100.3,108.1,164.0 16.5,57.4,86.7,120.1,139.0 29.2,55.8,97.2,129.6,155.9 15.9,35.7,70.9,108.4,164.6 22.2,38.1,74.5,115.6,151.8 3.0,38.4,82.1,110.0,157.7 23.6,59.2,75.7,125.4,151.1 17.9,38.6,99.8,131.6,153.1 15.6,58.2,101.2,107.7,141.4 33.8,45.4,70.1,118.3,154.6 21.7,36.0,93.4,117.2,167.0 4.4,62.6,88.2,122.7,149.8 25.0,51.9,80.0,111.3,163.7 20.5,52.2,76.8,122.3,147.6 21.4,61.7,91.4,118.9,163.2 5.7,42.6,81.0,106.7,169.6 22.8,41.1,84.8,126.7,153.4 6.9,35.1,88.5,113.4,146.5 7.2,53.8,99.4,117.5,143.6 7.5,39.4,67.7,114.6,152.2 9.6,41.4,74.8,119.5,162.0 10.5,56.4,72.0,121.7,167.5 12.8,54.2,74.2,131.9,167.8 18.3,46.4,87.4,135.4,146.1 31.2,40.5,82.5,133.4,149.0 32.4,62.3,71.2,136.0,158.9 2.8,64.5,91.1,119.8,144.5 26.6,62.9,76.6,121.0,162.9 14.3,40.1,100.5,135.1,150.5 28.4,48.1,84.1,137.7,169.3 8.9,51.6,100.8,114.2,138.8 30.6,44.0,78.6,117.7,154.3 6.7,56.7,85.2,133.1,140.8 28.8,61.2,95.2,108.7,156.5 12.5,49.6,101.5,124.0,150.8 11.9,36.5,80.7,137.5,145.5 17.5,53.3,67.4,107.3,145.8 32.9,37.4,97.6,106.4,164.3 11.1,34.9,81.4,113.8,140.5 16.8,61.4,103.0,125.7,161.4 9.9,42.9,78.9,127.6,143.9 20.3,50.1,96.1,109.6,160.2 24.4,42.4,87.1,118.0,142.7 19.1,41.7,75.3,137.1,150.1 10.8,48.7,66.8,126.4,141.7 25.3,65.2,71.6,129.1,166.4 24.8,52.5,79.6,113.0,151.5 4.7,55.5,69.1,130.1,165.5 6.0,59.5,96.7,110.4,140.2 29.6,60.3,68.1,134.2,138.5 23.1,48.4,78.2,121.4,141.1 12.2,38.9,86.3,105.6,168.4 30.9,49.1,93.0,104.9,148.5 13.1,55.1,73.3,112.5,159.6 0.5,58.9,69.8,123.1,144.8 31.5,50.8,90.2,134.0,163.5 9.3,57.1,76.1,110.8,165.2 13.9,54.7,96.4,112.0,139.6 32.7,47.8,68.3,133.7,158.0 7.8,36.8,68.6,124.7,166.7 15.2,43.4,92.1,124.9,162.3 1.1,44.3,73.7,126.9,154.0 27.0,63.9,92.7,132.5,159.2 2.1,65.7,98.8,130.6,169.0 20.0,44.7,85.8,128.8,143.0 30.2,40.8,73.0,128.3,164.9 28.1,60.9,77.6,124.3,141.9 15.0,56.1,98.2,136.4,157.1 5.3,36.2,102.1,126.0,139.3 8.5,49.9,99.0,111.1,168.2 17.2,59.7,92.4,131.3,160.5 14.6,45.8,80.3,123.7,166.1 19.5,47.4,102.6,132.2,157.4 18.7,39.8,94.6,111.8,155.0 33.2,60.6,72.3,106.0,152.8 27.9,61.9,67.1,105.3,165.8 6.3,53.6,70.4,120.4,142.3 10.3,42.0,104.3,130.4,148.8 16.3,57.9,93.7,116.2,167.2 25.8,60.1,83.7,129.9,143.3 34.1,37.1,94.9,116.7,168.7 8.2,45.1,77.2,109.0,149.5 32.1,49.3,103.7,115.2,146.7 3.6,66.0,103.4,134.8,145.2 1.8,64.8,83.3,134.5,158.6 2.5,65.5,94.3,123.4,147.0 24.1,46.1,77.8,128.0,161.1 31.8,58.4,84.5,132.8,152.6 23.4,50.4,94.0,116.0,148.2 1.5,53.0,95.6,119.2,161.6 5.0,43.1,66.5,130.9,158.4 21.1,63.6,72.6,129.4,160.8 26.2,63.3,91.8,136.7,147.9 34.5,51.1,89.0,117.0,156.8 11.5,58.7,101.9,114.9,139.9 0.8,51.3,69.4,118.6,162.6",
  "gpt": "224.0,445.3,596.8,828.2,962.8 28.8,317.6,566.9,896.7,1072.5 99.9,300.1,602.8,910.6,1047.3 181.9,319.8,559.7,837.9,1031.6 145.5,258.5,659.9,704.0,1105.9 120.2,392.3,585.6,784.4,917.5 193.1,379.5,644.9,848.2,1043.9 116.5,240.0,488.6,706.0,1112.9 154.7,261.2,510.9,748.0,1016.7 23.1,264.0,562.4,714.1,1057.2 161.6,407.0,518.5,813.9,1010.7 128.9,266.2,657.7,864.0,1027.6 114.7,397.6,666.0,702.1,940.8 225.8,309.6,483.6,769.8,1038.9 152.6,242.7,619.5,760.0,1129.7 30.6,434.2,591.5,796.2,1002.2 169.4,354.3,547.5,722.5,1103.5 143.6,357.2,526.8,794.4,986.3 149.9,427.4,607.1,774.3,1099.5 37.4,291.7,554.3,695.0,1148.6 156.3,280.4,577.9,822.0,1029.4 47.6,238.0,592.9,733.1,978.2 50.2,369.5,655.4,762.0,957.7 51.8,270.5,466.8,739.7,1018.7 66.3,283.3,513.2,778.7,1091.0 74.9,383.9,494.6,792.5,1134.3 93.9,371.8,508.9,866.3,1136.1 131.2,315.9,589.7,894.9,976.2 203.7,276.3,564.4,878.4,997.3 213.9,432.4,490.3,898.4,1065.8 20.5,447.4,605.1,780.9,964.6 176.6,435.8,524.2,788.1,1097.1 105.0,274.2,661.9,893.3,1006.2 188.3,325.8,573.6,908.2,1146.2 61.3,350.5,663.7,737.4,915.6 199.4,302.2,540.8,764.2,1036.1 45.7,386.5,579.9,876.4,936.5 191.2,422.8,631.8,708.0,1049.7 91.8,337.5,668.4,804.0,1008.0 86.0,248.0,551.9,905.9,972.0 127.3,364.2,464.7,697.3,974.1 218.6,256.3,646.7,692.8,1108.8 81.0,234.5,556.3,735.1,934.8 122.2,424.9,676.5,816.1,1086.2 68.8,294.1,542.5,830.0,959.7 141.7,340.9,635.9,711.7,1075.3 165.6,289.4,587.9,766.6,950.0 135.8,285.0,516.6,904.0,1004.3 77.8,332.9,460.5,820.2,942.4 171.2,451.1,492.6,842.7,1125.6 167.4,359.7,544.4,730.0,1014.8 32.1,377.0,477.4,851.9,1119.3 41.0,409.4,642.9,716.1,932.1 195.2,415.6,469.6,885.4,914.0 158.1,330.8,536.8,789.8,938.6 88.5,268.5,583.9,688.9,1140.1 201.7,334.6,617.8,685.3,992.5 97.3,375.1,503.9,728.2,1069.6 3.6,404.9,481.8,798.2,968.0 206.1,344.9,599.6,883.3,1101.2 63.5,388.8,521.7,718.1,1117.7 101.9,373.5,637.8,726.5,922.4 215.7,323.9,472.5,880.3,1059.1 53.6,250.3,474.8,809.4,1127.7 112.5,297.8,611.1,811.4,1093.2 9.9,303.8,506.2,824.7,1033.6 179.1,442.7,615.1,872.8,1067.9 16.3,455.4,652.1,856.0,1144.0 140.1,305.6,581.7,840.2,952.0 196.7,278.5,500.4,835.9,1115.7 185.7,421.2,531.1,807.0,944.1 110.1,381.3,649.2,900.6,1053.1 35.2,245.0,671.8,817.9,920.5 59.1,339.2,653.9,720.6,1138.5 124.0,411.3,613.1,861.1,1078.4 108.3,311.3,549.5,802.3,1124.1 138.3,321.6,673.8,869.8,1055.3 133.2,272.5,627.9,724.2,1041.6 222.0,417.8,496.9,690.6,1023.9 183.7,429.3,462.9,687.4,1122.2 43.6,366.6,486.0,786.4,947.8 72.2,287.0,683.5,853.7,995.3 118.6,394.7,621.1,753.6,1131.8 172.7,413.6,571.6,850.1,954.6 228.0,253.9,629.6,755.8,1141.9 56.9,307.6,529.0,709.8,1000.5 211.2,336.0,681.1,744.7,980.8 25.8,457.1,679.1,891.3,970.0 14.5,449.7,568.4,888.3,1063.6 18.7,453.4,625.8,800.0,983.1 163.7,313.5,534.5,832.8,1083.9 208.5,400.0,575.7,874.9,1021.1 159.5,342.9,622.8,750.4,990.8 12.6,361.9,633.6,776.3,1088.5 33.9,295.9,459.1,858.7,1061.5 147.6,439.8,499.0,845.1,1082.2 174.7,438.0,609.1,902.5,988.6 229.9,346.8,594.9,757.6,1051.2 83.7,402.1,670.2,742.1,929.7 7.1,348.4,479.8,771.5,1095.3",
  "gpt_reason": "435.4,871.7,1175.4,1647.7,1897.7 51.7,627.0,1109.4,1776.4,2106.8 183.9,586.5,1183.9,1807.4,2052.7 350.2,631.7,1096.5,1664.6,2026.8 272.9,498.8,1304.6,1395.6,2168.5 224.6,769.2,1149.8,1554.8,1821.8 371.7,748.4,1272.8,1684.5,2047.8 215.9,462.8,967.4,1401.0,2177.7 289.9,503.3,1012.9,1489.7,1999.6 42.6,508.5,1100.4,1417.5,2072.8 308.7,795.0,1026.0,1621.9,1991.5 240.6,513.0,1299.8,1715.3,2017.9 211.2,778.7,1317.1,1391.0,1858.1 439.0,609.5,957.8,1528.3,2039.6 286.2,467.2,1221.4,1511.3,2213.0 56.3,849.7,1162.9,1584.2,1974.2 325.9,699.5,1073.9,1434.9,2163.2 268.7,704.5,1039.0,1579.4,1940.5 281.5,837.2,1192.5,1537.5,2154.9 74.2,567.1,1087.2,1382.0,2251.6 294.7,544.1,1131.6,1639.8,2021.9 92.1,457.7,1167.2,1459.6,1928.1 97.0,726.2,1294.8,1515.7,1888.9 102.1,522.3,926.9,1472.1,2004.3 130.9,548.9,1016.9,1546.2,2137.1 144.5,756.8,980.5,1573.1,2220.8 174.2,729.9,1008.7,1719.1,2225.1 244.7,623.0,1157.9,1772.4,1923.6 394.9,535.8,1105.2,1739.7,1964.9 413.2,845.7,971.6,1780.7,2093.9 38.2,876.5,1188.2,1550.6,1902.2 341.3,854.6,1034.9,1563.9,2150.0 192.7,531.2,1308.8,1767.4,1983.4 364.1,644.1,1123.3,1802.7,2246.7 119.3,695.7,1313.2,1467.7,1817.2 385.6,591.1,1061.0,1520.0,2035.1 86.9,760.2,1136.2,1734.8,1850.3 367.9,827.3,1249.8,1406.0,2056.3 170.5,667.1,1321.0,1602.7,1987.0 162.0,476.7,1082.7,1798.1,1915.4 236.4,718.0,922.3,1387.0,1919.3 425.7,491.8,1277.5,1377.2,2173.1 153.4,453.3,1092.0,1463.6,1846.0 228.5,831.5,1338.6,1626.2,2128.0 135.0,572.4,1065.7,1651.7,1893.6 265.4,675.6,1258.5,1414.2,2111.0 316.7,562.4,1153.9,1524.4,1875.8 252.4,554.3,1021.6,1794.3,1979.6 148.7,653.3,912.1,1635.8,1862.8 329.9,886.0,976.3,1675.0,2204.7 321.4,709.1,1070.6,1455.3,1996.0 60.2,744.0,945.0,1693.3,2191.3 78.8,799.7,1267.7,1421.9,1841.7 376.2,813.5,931.1,1753.4,1813.0 298.6,648.7,1056.7,1568.3,1854.8 166.2,517.3,1145.5,1367.2,2232.6 389.9,658.3,1216.0,1357.8,1955.3 178.6,739.8,998.6,1450.7,2103.2 4.7,790.8,954.1,1588.2,1906.7 399.1,683.9,1179.7,1748.5,2158.7 125.9,765.1,1030.7,1426.4,2187.6 188.5,734.9,1262.8,1444.1,1830.7 418.7,639.9,935.9,1744.4,2077.0 106.8,481.9,940.3,1612.6,2208.7 207.2,582.3,1202.0,1616.8,2140.7 13.8,596.2,1003.1,1644.2,2031.1 346.1,867.4,1211.0,1726.5,2098.4 28.2,893.9,1286.4,1701.9,2242.2 261.3,601.1,1140.4,1670.5,1881.2 381.7,540.0,993.9,1660.3,2182.3 359.5,822.9,1048.1,1607.7,1867.2 202.2,752.8,1282.3,1785.6,2064.2 69.2,472.3,1329.9,1631.1,1826.0 115.0,671.3,1290.7,1431.1,2228.6 232.8,804.7,1206.2,1710.4,2115.2 197.3,613.4,1078.3,1597.4,2200.9 257.0,636.0,1335.3,1722.7,2068.9 248.2,527.6,1241.3,1439.6,2043.8 431.4,818.6,984.6,1371.7,2013.5 354.9,841.7,916.6,1362.4,2195.6 82.7,721.8,962.4,1558.8,1871.1 140.3,558.2,1351.8,1697.6,1959.2 220.9,773.9,1227.1,1499.2,2217.1 333.6,808.8,1119.1,1688.9,1884.7 444.0,486.7,1245.9,1503.5,2237.5 111.1,606.0,1043.4,1410.2,1969.5 408.9,662.7,1347.4,1481.8,1932.3 48.0,898.1,1343.5,1763.8,1910.9 23.4,882.3,1114.7,1760.2,2088.9 33.6,889.8,1235.9,1592.6,1936.1 312.5,619.0,1053.1,1655.7,2124.1 403.9,782.3,1127.7,1730.5,2008.2 303.0,680.0,1231.5,1494.3,1949.5 18.8,713.5,1254.2,1541.5,2131.6 64.4,577.6,907.9,1706.5,2084.6 277.4,862.3,989.3,1679.8,2120.2 336.9,858.2,1197.7,1789.9,1944.9 448.2,688.3,1171.9,1507.5,2060.7 157.7,786.7,1325.5,1477.9,1837.3 9.6,692.3,949.8,1533.0,2145.8"
 }
}