{
 "video": "https://modelfatigue.news/v/who-beats-jev/",
 "numbers": [
  {
   "id": "n-test",
   "what": "Banking77 test messages every model answered",
   "value": 3080,
   "shown": "3,080",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-options",
   "what": "Options in each question: the 9 nearest reasons plus none of these",
   "value": 10,
   "shown": "10",
   "whose": "Model Fatigue",
   "source": "Our protocol addendum for the 10-option question, fixed before any test message was sent",
   "url": null,
   "read": "2026-10-05T21:38",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-near",
   "what": "Reasons on each question's list, besides none of these",
   "value": 9,
   "shown": "9",
   "whose": "Model Fatigue",
   "source": "Our protocol addendum for the 10-option question, fixed before any test message was sent",
   "url": null,
   "read": "2026-10-05T21:38",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-full",
   "what": "Options in the full question (77 reasons plus none of these)",
   "value": 78,
   "shown": "78",
   "whose": "Model Fatigue",
   "source": "Our results write-up (RESULTS.md, section \"Who beats Jev now?\")",
   "url": null,
   "read": "2026-10-06",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gold-on",
   "what": "Messages whose right reason is on its ten-option list",
   "value": 3073,
   "shown": "3,073",
   "whose": "Model Fatigue",
   "source": "Our results write-up (RESULTS.md, section \"Who beats Jev now?\")",
   "url": null,
   "read": "2026-10-06",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-78",
   "what": "Jev on the full 78-option question, the same day (its own pass)",
   "value": 0.9392857142857143,
   "shown": "93.9%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-acc",
   "what": "Quyet-1.0-Large, share right",
   "value": 0.9448051948051948,
   "shown": "94.5%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-acc-lo",
   "what": "Quyet-1.0-Large, share right, 95% interval, low",
   "value": 0.9361721566889261,
   "shown": "93.6",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-acc-hi",
   "what": "Quyet-1.0-Large, share right, 95% interval, high",
   "value": 0.9523300283779073,
   "shown": "95.2",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-correct",
   "what": "Quyet-1.0-Large, messages sorted correctly",
   "value": 2910,
   "shown": "2,910",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-none",
   "what": "Quyet-1.0-Large, times it answered none of these",
   "value": 10,
   "shown": "10",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-p50",
   "what": "Quyet-1.0-Large, median time per request from Berlin",
   "value": 0.431315,
   "shown": "431 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-p95",
   "what": "Quyet-1.0-Large, 95th percentile time per request from Berlin",
   "value": 0.4589674999999999,
   "shown": "459 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-p99",
   "what": "Quyet-1.0-Large, 99th percentile time per request from Berlin",
   "value": 0.5803487000000002,
   "shown": "580 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-per1k",
   "what": "Quyet-1.0-Large, dollars per 1,000 decisions (our GPU time while the messages ran, as Modal billed it, one request at a time)",
   "value": 0.6024347897956766,
   "shown": "$0.602",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-ece",
   "what": "Quyet-1.0-Large, calibration error of its probabilities (ECE; lower is closer)",
   "value": 0.018414480519480466,
   "shown": "0.018",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-auroc",
   "what": "Quyet-1.0-Large, how well its probabilities separate its right answers from its wrong ones (AUROC)",
   "value": 0.9015716595916717,
   "shown": "0.90",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-thr",
   "what": "Quyet-1.0-Large, cut-off on its top probability, set on the separate set",
   "value": 0.88,
   "shown": "0.880",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-cov",
   "what": "Quyet-1.0-Large, share of messages it handled without a person",
   "value": 0.9159090909090909,
   "shown": "91.6%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-risk",
   "what": "Quyet-1.0-Large, wrong among the messages it handled without a person",
   "value": 0.02445941155618575,
   "shown": "2.45%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-acc",
   "what": "deck-31B, share right",
   "value": 0.9396103896103896,
   "shown": "94.0%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-acc-lo",
   "what": "deck-31B, share right, 95% interval, low",
   "value": 0.930637477333861,
   "shown": "93.1",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-acc-hi",
   "what": "deck-31B, share right, 95% interval, high",
   "value": 0.9474880398781773,
   "shown": "94.7",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-correct",
   "what": "deck-31B, messages sorted correctly",
   "value": 2894,
   "shown": "2,894",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-none",
   "what": "deck-31B, times it answered none of these",
   "value": 24,
   "shown": "24",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-p50",
   "what": "deck-31B, median time per request from Berlin",
   "value": 0.53933,
   "shown": "539 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-p95",
   "what": "deck-31B, 95th percentile time per request from Berlin",
   "value": 0.8446684999999994,
   "shown": "845 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-p99",
   "what": "deck-31B, 99th percentile time per request from Berlin",
   "value": 1.1863314000000003,
   "shown": "1,186 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-per1k",
   "what": "deck-31B, dollars per 1,000 decisions (our GPU time while the messages ran, as Modal billed it, one request at a time)",
   "value": 0.8173932728847128,
   "shown": "$0.817",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-ece",
   "what": "deck-31B, calibration error of its probabilities (ECE; lower is closer)",
   "value": 0.0397056662255735,
   "shown": "0.040",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-auroc",
   "what": "deck-31B, how well its probabilities separate its right answers from its wrong ones (AUROC)",
   "value": 0.7076859055814403,
   "shown": "0.71",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-thr",
   "what": "deck-31B, cut-off on its top probability, set on the separate set",
   "value": 0.9961795824130063,
   "shown": "0.996",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-cov",
   "what": "deck-31B, share of messages it handled without a person",
   "value": 0.012337662337662338,
   "shown": "1.2%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-risk",
   "what": "deck-31B, wrong among the messages it handled without a person",
   "value": 0.02631578947368421,
   "shown": "2.63%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-acc",
   "what": "Jev, share right",
   "value": 0.938961038961039,
   "shown": "93.9%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-acc-lo",
   "what": "Jev, share right, 95% interval, low",
   "value": 0.9299469172345147,
   "shown": "93.0",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-acc-hi",
   "what": "Jev, share right, 95% interval, high",
   "value": 0.9468815164956743,
   "shown": "94.7",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-correct",
   "what": "Jev, messages sorted correctly",
   "value": 2892,
   "shown": "2,892",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-none",
   "what": "Jev, times it answered none of these",
   "value": 19,
   "shown": "19",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-p50",
   "what": "Jev, median time per request from Berlin",
   "value": 0.22939,
   "shown": "229 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-p95",
   "what": "Jev, 95th percentile time per request from Berlin",
   "value": 0.28161400000000003,
   "shown": "282 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-p99",
   "what": "Jev, 99th percentile time per request from Berlin",
   "value": 0.3483808,
   "shown": "348 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-per1k",
   "what": "Jev, dollars per 1,000 decisions (price list × input tokens counted)",
   "value": 0.02608074545454546,
   "shown": "$0.026",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-ece",
   "what": "Jev, calibration error of its probabilities (ECE; lower is closer)",
   "value": 0.03410064935064942,
   "shown": "0.034",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-auroc",
   "what": "Jev, how well its probabilities separate its right answers from its wrong ones (AUROC)",
   "value": 0.8419466025131691,
   "shown": "0.84",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-thr",
   "what": "Jev, cut-off on its top probability, set on the separate set",
   "value": 0.99,
   "shown": "0.990",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-cov",
   "what": "Jev, share of messages it handled without a person",
   "value": 0.837987012987013,
   "shown": "83.8%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-risk",
   "what": "Jev, wrong among the messages it handled without a person",
   "value": 0.019372336303758234,
   "shown": "1.94%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-acc",
   "what": "Gemma-4-31B-it through Quyet's prompt, share right",
   "value": 0.9376623376623376,
   "shown": "93.8%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-acc-lo",
   "what": "Gemma-4-31B-it through Quyet's prompt, share right, 95% interval, low",
   "value": 0.9285666005105627,
   "shown": "92.9",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-acc-hi",
   "what": "Gemma-4-31B-it through Quyet's prompt, share right, 95% interval, high",
   "value": 0.9456676662559275,
   "shown": "94.6",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-correct",
   "what": "Gemma-4-31B-it through Quyet's prompt, messages sorted correctly",
   "value": 2888,
   "shown": "2,888",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-none",
   "what": "Gemma-4-31B-it through Quyet's prompt, times it answered none of these",
   "value": 31,
   "shown": "31",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-p50",
   "what": "Gemma-4-31B-it through Quyet's prompt, median time per request from Berlin",
   "value": 0.46102,
   "shown": "461 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-p95",
   "what": "Gemma-4-31B-it through Quyet's prompt, 95th percentile time per request from Berlin",
   "value": 0.49561799999999995,
   "shown": "496 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-p99",
   "what": "Gemma-4-31B-it through Quyet's prompt, 99th percentile time per request from Berlin",
   "value": 0.8306614000000004,
   "shown": "831 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-per1k",
   "what": "Gemma-4-31B-it through Quyet's prompt, dollars per 1,000 decisions (our GPU time while the messages ran, as Modal billed it, one request at a time)",
   "value": 0.6576233190632715,
   "shown": "$0.658",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-ece",
   "what": "Gemma-4-31B-it through Quyet's prompt, calibration error of its probabilities (ECE; lower is closer)",
   "value": 0.04746237012987019,
   "shown": "0.047",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-auroc",
   "what": "Gemma-4-31B-it through Quyet's prompt, how well its probabilities separate its right answers from its wrong ones (AUROC)",
   "value": 0.7921734331717452,
   "shown": "0.79",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-thr",
   "what": "Gemma-4-31B-it through Quyet's prompt, cut-off on its top probability, set on the separate set",
   "value": 0.9988,
   "shown": "0.999",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-cov",
   "what": "Gemma-4-31B-it through Quyet's prompt, share of messages it handled without a person",
   "value": 0.6396103896103896,
   "shown": "64.0%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-risk",
   "what": "Gemma-4-31B-it through Quyet's prompt, wrong among the messages it handled without a person",
   "value": 0.024873096446700507,
   "shown": "2.49%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-acc",
   "what": "Clef-Flash, share right",
   "value": 0.952922077922078,
   "shown": "95.3%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-acc-lo",
   "what": "Clef-Flash, share right, 95% interval, low",
   "value": 0.9448609801002017,
   "shown": "94.5",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-acc-hi",
   "what": "Clef-Flash, share right, 95% interval, high",
   "value": 0.9598547484897493,
   "shown": "96.0",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-correct",
   "what": "Clef-Flash, messages sorted correctly",
   "value": 2935,
   "shown": "2,935",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-none",
   "what": "Clef-Flash, times it answered none of these",
   "value": 1,
   "shown": "1",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-p50",
   "what": "Clef-Flash, median time per request from Berlin",
   "value": 0.28996,
   "shown": "290 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-p95",
   "what": "Clef-Flash, 95th percentile time per request from Berlin",
   "value": 0.8861869999999998,
   "shown": "886 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-p99",
   "what": "Clef-Flash, 99th percentile time per request from Berlin",
   "value": 1.5200863000000011,
   "shown": "1,520 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-per1k",
   "what": "Clef-Flash, dollars per 1,000 decisions (price list × input tokens counted)",
   "value": 0.042303126623376625,
   "shown": "$0.042",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-ece",
   "what": "Clef-Flash, calibration error of its probabilities (ECE; lower is closer)",
   "value": 0.04607756493506499,
   "shown": "0.046",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-auroc",
   "what": "Clef-Flash, how well its probabilities separate its right answers from its wrong ones (AUROC)",
   "value": 0.862645832109499,
   "shown": "0.86",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-thr",
   "what": "Clef-Flash, cut-off on its top probability, set on the separate set",
   "value": 0.8,
   "shown": "0.800",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-cov",
   "what": "Clef-Flash, share of messages it handled without a person",
   "value": 0.9324675324675324,
   "shown": "93.2%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-risk",
   "what": "Clef-Flash, wrong among the messages it handled without a person",
   "value": 0.027855153203342618,
   "shown": "2.79%",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-vj-ar",
   "what": "Messages Quyet-1.0-Large got right and Jev got wrong",
   "value": 36,
   "shown": "36",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-vj-jr",
   "what": "Messages Jev got right and Quyet-1.0-Large got wrong",
   "value": 18,
   "shown": "18",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-vj-ar",
   "what": "Messages deck-31B got right and Jev got wrong",
   "value": 23,
   "shown": "23",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-vj-jr",
   "what": "Messages Jev got right and deck-31B got wrong",
   "value": 21,
   "shown": "21",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-vj-ar",
   "what": "Messages Gemma-4-31B-it through Quyet's prompt got right and Jev got wrong",
   "value": 21,
   "shown": "21",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-vj-jr",
   "what": "Messages Jev got right and Gemma-4-31B-it through Quyet's prompt got wrong",
   "value": 25,
   "shown": "25",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-vj-ar",
   "what": "Messages Clef-Flash got right and Jev got wrong",
   "value": 60,
   "shown": "60",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-vj-jr",
   "what": "Messages Jev got right and Clef-Flash got wrong",
   "value": 17,
   "shown": "17",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "price-jev",
   "what": "Jev, price per million input tokens",
   "value": 0.042,
   "shown": "$0.042",
   "whose": "TypeSafe",
   "source": "TypeSafe's models page (Jev's price)",
   "url": "https://docs.typesafe.ai/models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "q-billed",
   "what": "Everything Modal billed for Quyet's GPU app",
   "value": 4.348871539999999,
   "shown": "$4.35",
   "whose": "Model Fatigue",
   "source": "Modal's bill for our GPU apps, read back per pass (wbj_gpu_cost.json)",
   "url": null,
   "read": "2026-10-06T01:05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "q-overhead",
   "what": "Of that, start-up, warm-ups and idle time",
   "value": 2.3016547591485526,
   "shown": "$2.30",
   "whose": "Model Fatigue",
   "source": "Modal's bill for our GPU apps, read back per pass (wbj_gpu_cost.json)",
   "url": null,
   "read": "2026-10-06T01:05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "q-calls",
   "what": "Calls Quyet answered in all (the separate set, the test and a repeat pass)",
   "value": 3853,
   "shown": "3,853",
   "whose": "Model Fatigue",
   "source": "Modal's bill for our GPU apps, read back per pass (wbj_gpu_cost.json)",
   "url": null,
   "read": "2026-10-06T01:05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "q-whole",
   "what": "Quyet's whole bill spread over every call it answered, dollars per 1,000 (our arithmetic)",
   "value": 1.1286975188165065,
   "shown": "$1.13",
   "whose": "Model Fatigue",
   "source": "Modal's bill for our GPU apps, read back per pass (wbj_gpu_cost.json)",
   "url": null,
   "read": "2026-10-06T01:05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "q-basis",
   "what": "Quyet per 1,000 on JevBench's basis: OpenRouter's list price for plain Gemma × the tokens of our run (our arithmetic)",
   "value": 0.02424816233766234,
   "shown": "$0.024",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "dev-n",
   "what": "Messages in the separate set every cut-off was set on",
   "value": 573,
   "shown": "573",
   "whose": "Model Fatigue",
   "source": "Our pass over the separate set of 573 messages the cut-offs were set on (dev records, our arithmetic in derived.json)",
   "url": null,
   "read": "2026-10-05T20:36",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-top",
   "what": "deck-31B's most confident answers on the separate set that were all right before the first wrong one",
   "value": 9,
   "shown": "9",
   "whose": "Model Fatigue",
   "source": "Our pass over the separate set of 573 messages the cut-offs were set on (dev records, our arithmetic in derived.json)",
   "url": null,
   "read": "2026-10-05T20:36",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-floor",
   "what": "deck-31B, the lowest error rate any cut-off below that point reaches on the separate set",
   "value": 0.02681992337164751,
   "shown": "2.7%",
   "whose": "Model Fatigue",
   "source": "Our pass over the separate set of 573 messages the cut-offs were set on (dev records, our arithmetic in derived.json)",
   "url": null,
   "read": "2026-10-05T20:36",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "target",
   "what": "Wrong answers allowed among those a model handles alone, the target every cut-off was set for",
   "value": 0.02,
   "shown": "2%",
   "whose": "Model Fatigue",
   "source": "Our protocol addendum for the 10-option question, fixed before any test message was sent",
   "url": null,
   "read": "2026-10-05T21:38",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-ranked",
   "what": "Systems JevBench v1.6.0 ranks",
   "value": 92,
   "shown": "92",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-cap",
   "what": "Quyet-1.0-Large, JevBench Capability",
   "value": 81.68880454670384,
   "shown": "81.7",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-cap-lo",
   "what": "Quyet-1.0-Large, JevBench Capability, 95% interval, low",
   "value": 77.9843219931321,
   "shown": "78.0",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-cap-hi",
   "what": "Quyet-1.0-Large, JevBench Capability, 95% interval, high",
   "value": 83.55754626687737,
   "shown": "83.6",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-int",
   "what": "Quyet-1.0-Large, JevBench's right-answers part (Intelligence)",
   "value": 73.40135940607429,
   "shown": "73.4",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-cal",
   "what": "Quyet-1.0-Large, JevBench's confidence part (Calibration)",
   "value": 89.97624968733336,
   "shown": "90.0",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-rank",
   "what": "Quyet-1.0-Large, place among Jev-class systems on Capability",
   "value": 1,
   "shown": "1",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-off",
   "what": "Quyet-1.0-Large, place on JevBench's official ranking",
   "value": 1,
   "shown": "1",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-items",
   "what": "Quyet-1.0-Large, JevBench questions it answered",
   "value": 1500,
   "shown": "1,500",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-choice",
   "what": "Quyet-1.0-Large, share right on JevBench's multiple-choice questions",
   "value": 0.8706666666666667,
   "shown": "87%",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-deck-cap",
   "what": "deck-31B, JevBench Capability",
   "value": 77.5907429125261,
   "shown": "77.6",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-deck-cap-lo",
   "what": "deck-31B, JevBench Capability, 95% interval, low",
   "value": 73.77453269228936,
   "shown": "73.8",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-deck-cap-hi",
   "what": "deck-31B, JevBench Capability, 95% interval, high",
   "value": 80.24074177369353,
   "shown": "80.2",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-deck-int",
   "what": "deck-31B, JevBench's right-answers part (Intelligence)",
   "value": 73.02512554551848,
   "shown": "73.0",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-deck-cal",
   "what": "deck-31B, JevBench's confidence part (Calibration)",
   "value": 82.15636027953371,
   "shown": "82.2",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-deck-rank",
   "what": "deck-31B, place among Jev-class systems on Capability",
   "value": 2,
   "shown": "2",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-deck-off",
   "what": "deck-31B, place on JevBench's official ranking",
   "value": 5,
   "shown": "5",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-deck-items",
   "what": "deck-31B, JevBench questions it answered",
   "value": 1500,
   "shown": "1,500",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-deck-choice",
   "what": "deck-31B, share right on JevBench's multiple-choice questions",
   "value": 0.86,
   "shown": "86%",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-jev-cap",
   "what": "Jev, JevBench Capability",
   "value": 76.48401425422394,
   "shown": "76.5",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-jev-cap-lo",
   "what": "Jev, JevBench Capability, 95% interval, low",
   "value": 69.1907647929543,
   "shown": "69.2",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-jev-cap-hi",
   "what": "Jev, JevBench Capability, 95% interval, high",
   "value": 78.7556482385829,
   "shown": "78.8",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-jev-int",
   "what": "Jev, JevBench's right-answers part (Intelligence)",
   "value": 62.40124449091305,
   "shown": "62.4",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-jev-cal",
   "what": "Jev, JevBench's confidence part (Calibration)",
   "value": 90.56678401753481,
   "shown": "90.6",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-jev-rank",
   "what": "Jev, place among Jev-class systems on Capability",
   "value": 3,
   "shown": "3",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-jev-off",
   "what": "Jev, place on JevBench's official ranking",
   "value": 2,
   "shown": "2",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-jev-items",
   "what": "Jev, JevBench questions it answered",
   "value": 600,
   "shown": "600",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-jev-choice",
   "what": "Jev, share right on JevBench's multiple-choice questions",
   "value": 0.8166666666666667,
   "shown": "82%",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-cap",
   "what": "Clef-Flash, JevBench Capability",
   "value": 63.824241207952895,
   "shown": "63.8",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-cap-lo",
   "what": "Clef-Flash, JevBench Capability, 95% interval, low",
   "value": 59.38252934298857,
   "shown": "59.4",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-cap-hi",
   "what": "Clef-Flash, JevBench Capability, 95% interval, high",
   "value": 65.53439425442862,
   "shown": "65.5",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-int",
   "what": "Clef-Flash, JevBench's right-answers part (Intelligence)",
   "value": 42.18368888957246,
   "shown": "42.2",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-cal",
   "what": "Clef-Flash, JevBench's confidence part (Calibration)",
   "value": 85.46479352633332,
   "shown": "85.5",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-rank",
   "what": "Clef-Flash, place among Jev-class systems on Capability",
   "value": 11,
   "shown": "11",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-off",
   "what": "Clef-Flash, place on JevBench's official ranking",
   "value": 21,
   "shown": "21",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-items",
   "what": "Clef-Flash, JevBench questions it answered",
   "value": 1500,
   "shown": "1,500",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-choice",
   "what": "Clef-Flash, share right on JevBench's multiple-choice questions",
   "value": 0.6666666666666666,
   "shown": "67%",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-usd",
   "what": "Quyet, JevBench's price per 1,000 decisions (its estimate)",
   "value": 0.04451059729064039,
   "shown": "$0.045",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-p50",
   "what": "Quyet, JevBench's median time (after its adjustment)",
   "value": 0.3815564611926675,
   "shown": "382 ms",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-quyet-raw",
   "what": "Quyet, the median time JevBench measured",
   "value": 0.11577823059633374,
   "shown": "116 ms",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-sage-off",
   "what": "Sage 1.3.0, place on JevBench v1.6.0's official ranking",
   "value": 9,
   "shown": "9",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-sage-cap",
   "what": "Sage 1.3.0, JevBench v1.6.0 Capability (outside the Jev-class cost cap there)",
   "value": 77.08747420427488,
   "shown": "77.1",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-recheck",
   "what": "Jev on JevBench's second check of v1.6.0, on new questions",
   "value": 77.5,
   "shown": "77.5",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench's page text: method notes, caps, the latency adjustment, the Jev re-check",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-sage-cap",
   "what": "Sage 1.3.0, Capability on JevBench v1.6.1",
   "value": 78.59554275566781,
   "shown": "78.6",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-sage-off",
   "what": "Sage 1.3.0, place on JevBench v1.6.1's official ranking",
   "value": 1,
   "shown": "1",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-sage-rank",
   "what": "Sage 1.3.0, place among Jev-class systems on v1.6.1's Capability",
   "value": 2,
   "shown": "2",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-sage-items",
   "what": "Sage 1.3.0, JevBench v1.6.1 questions it answered",
   "value": 1500,
   "shown": "1,500",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-quyet-cap",
   "what": "Quyet-1.0-Large, Capability on JevBench v1.6.1",
   "value": 81.68880454670384,
   "shown": "81.7",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-quyet-off",
   "what": "Quyet-1.0-Large, place on JevBench v1.6.1's official ranking",
   "value": 3,
   "shown": "3",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-quyet-rank",
   "what": "Quyet-1.0-Large, place among Jev-class systems on v1.6.1's Capability",
   "value": 1,
   "shown": "1",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-quyet-items",
   "what": "Quyet-1.0-Large, JevBench v1.6.1 questions it answered",
   "value": 1500,
   "shown": "1,500",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-jev-cap",
   "what": "Jev, Capability on JevBench v1.6.1",
   "value": 77.12647993588574,
   "shown": "77.1",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-jev-off",
   "what": "Jev, place on JevBench v1.6.1's official ranking",
   "value": 2,
   "shown": "2",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-jev-rank",
   "what": "Jev, place among Jev-class systems on v1.6.1's Capability",
   "value": 4,
   "shown": "4",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-jev-items",
   "what": "Jev, JevBench v1.6.1 questions it answered",
   "value": 1500,
   "shown": "1,500",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-deck-cap",
   "what": "deck-31B, Capability on JevBench v1.6.1",
   "value": 77.5907429125261,
   "shown": "77.6",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-deck-off",
   "what": "deck-31B, place on JevBench v1.6.1's official ranking",
   "value": 7,
   "shown": "7",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-deck-rank",
   "what": "deck-31B, place among Jev-class systems on v1.6.1's Capability",
   "value": 3,
   "shown": "3",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-deck-items",
   "what": "deck-31B, JevBench v1.6.1 questions it answered",
   "value": 1500,
   "shown": "1,500",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-flash-cap",
   "what": "Clef-Flash, Capability on JevBench v1.6.1",
   "value": 63.824241207952895,
   "shown": "63.8",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-flash-off",
   "what": "Clef-Flash, place on JevBench v1.6.1's official ranking",
   "value": 21,
   "shown": "21",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-flash-rank",
   "what": "Clef-Flash, place among Jev-class systems on v1.6.1's Capability",
   "value": 12,
   "shown": "12",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "n-flash-items",
   "what": "Clef-Flash, JevBench v1.6.1 questions it answered",
   "value": 1500,
   "shown": "1,500",
   "whose": "JevBench (Benchmark Heaven)",
   "source": "JevBench v1.6.1, the later version (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models",
   "read": "2026-10-06T11:00",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gold-off",
   "what": "Messages whose right reason is not on the list (none of these counts as right)",
   "value": 7.0,
   "shown": "7",
   "whose": "Model Fatigue, arithmetic on our run",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-cents",
   "what": "Quyet-1.0-Large, cents per 1,000 decisions (our arithmetic)",
   "value": 60.243478979567655,
   "shown": "60.2",
   "whose": "Model Fatigue, arithmetic on our run",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-cents",
   "what": "deck-31B, cents per 1,000 decisions (our arithmetic)",
   "value": 81.73932728847127,
   "shown": "81.7",
   "whose": "Model Fatigue, arithmetic on our run",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "jev-cents",
   "what": "Jev, cents per 1,000 decisions (our arithmetic)",
   "value": 2.6080745454545458,
   "shown": "2.6",
   "whose": "Model Fatigue, arithmetic on our run",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-cents",
   "what": "Gemma-4-31B-it through Quyet's prompt, cents per 1,000 decisions (our arithmetic)",
   "value": 65.76233190632715,
   "shown": "65.8",
   "whose": "Model Fatigue, arithmetic on our run",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-cents",
   "what": "Clef-Flash, cents per 1,000 decisions (our arithmetic)",
   "value": 4.230312662337663,
   "shown": "4.2",
   "whose": "Model Fatigue, arithmetic on our run",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-model",
   "what": "Quyet-1.0-Large, median time the model itself took on our GPU (its server's count)",
   "value": 0.06565,
   "shown": "66 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-model",
   "what": "Gemma-4-31B-it through Quyet's prompt, median time the model itself took on our GPU (its server's count)",
   "value": 0.067515,
   "shown": "68 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-model",
   "what": "deck-31B, median time the model itself took on our GPU (its server's count)",
   "value": 0.16881000000000002,
   "shown": "169 ms",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-vj",
   "what": "Quyet-1.0-Large minus Jev, share right, same messages",
   "value": 0.5844155844155844,
   "shown": "+0.58 points",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-vj-lo",
   "what": "Quyet-1.0-Large minus Jev, 95% interval, low",
   "value": 0.12987012987012986,
   "shown": "+0.13",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "quyet-vj-hi",
   "what": "Quyet-1.0-Large minus Jev, 95% interval, high",
   "value": 1.0714285714285714,
   "shown": "+1.07",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-vj",
   "what": "deck-31B minus Jev, share right, same messages",
   "value": 0.06493506493506493,
   "shown": "+0.06 points",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-vj-lo",
   "what": "deck-31B minus Jev, 95% interval, low",
   "value": -0.35714285714285715,
   "shown": "−0.36",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "deck-vj-hi",
   "what": "deck-31B minus Jev, 95% interval, high",
   "value": 0.487012987012987,
   "shown": "+0.49",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-vj",
   "what": "Gemma-4-31B-it through Quyet's prompt minus Jev, share right, same messages",
   "value": -0.12987012987012986,
   "shown": "−0.13 points",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-vj-lo",
   "what": "Gemma-4-31B-it through Quyet's prompt minus Jev, 95% interval, low",
   "value": -0.551948051948052,
   "shown": "−0.55",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "gemma-vj-hi",
   "what": "Gemma-4-31B-it through Quyet's prompt minus Jev, 95% interval, high",
   "value": 0.2922077922077922,
   "shown": "+0.29",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-vj",
   "what": "Clef-Flash minus Jev, share right, same messages",
   "value": 1.396103896103896,
   "shown": "+1.40 points",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-vj-lo",
   "what": "Clef-Flash minus Jev, 95% interval, low",
   "value": 0.8441558441558441,
   "shown": "+0.84",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "flash-vj-hi",
   "what": "Clef-Flash minus Jev, 95% interval, high",
   "value": 1.9805194805194806,
   "shown": "+1.98",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "train",
   "what": "Quyet minus plain Gemma through Quyet's prompt: what the training is worth",
   "value": 0.7142857142857143,
   "shown": "+0.71 points",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "train-lo",
   "what": "The training's worth, 95% interval, low",
   "value": 0.2597402597402597,
   "shown": "+0.26",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "train-hi",
   "what": "The training's worth, 95% interval, high",
   "value": 1.2012987012987013,
   "shown": "+1.20",
   "whose": "Model Fatigue",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "ph",
   "what": "Clef-Flash minus Quyet, share right (decided after the run: post hoc)",
   "value": 0.8116883116883189,
   "shown": "+0.81 points",
   "whose": "Model Fatigue, arithmetic on our run",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "alone-more",
   "what": "Quyet minus plain Gemma: more messages in 100 handled without a person (our arithmetic)",
   "value": 27.629870129870127,
   "shown": "+28",
   "whose": "Model Fatigue, arithmetic on our run",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "r-quyet",
   "what": "Quyet's cost per decision as we ran it against Jev's (our arithmetic)",
   "value": 23.09883323103718,
   "shown": "23×",
   "whose": "Model Fatigue, arithmetic on our run",
   "source": "Our run of the 10-option question: Jev, Clef-Flash, Quyet-1.0-Large, deck-31B and plain Gemma",
   "url": null,
   "read": "2026-10-05",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-flash-gap",
   "what": "Quyet minus Clef-Flash on JevBench Capability, rounded (our arithmetic)",
   "value": 17.86456333875094,
   "shown": "18",
   "whose": "Model Fatigue, arithmetic on JevBench's numbers",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "b-gap",
   "what": "Quyet minus Jev on JevBench Capability (our arithmetic)",
   "value": 5.204790292479899,
   "shown": "5.2",
   "whose": "Model Fatigue, arithmetic on JevBench's numbers",
   "source": "JevBench v1.6.0, Benchmark Heaven's leaderboard of decision models (its results JSON)",
   "url": "https://benchmarkheaven.com/jev-models/v1.6.0",
   "read": "2026-10-05T20:18",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  }
 ]
}