{
 "schema": 1,
 "date": "2026-10-25",
 "edition": "2026.10",
 "machine": "rtx-5090",
 "device": "cuda",
 "suite": "public-benchmark",
 "servers": {
  "ollaya": "0.18.0"
 },
 "protocol": {
  "benchmark": "Bespoke Labs' public decision benchmark: 13 subsets of 11 human-labeled public datasets, 3,880 questions",
  "source": "https://github.com/bespokelabsai/nimble/blob/main/docs/PUBLIC_BENCHMARKS.md",
  "runner": "Bespoke's own evaluate_public_jev, through convert/ollaya_convert/bench_public.py",
  "client": "one request (one question) at a time over HTTP to /v1/systemone, timed by the client",
  "accuracy": "macro: the mean of the 13 subsets; a rejected request counts as wrong",
  "ece": "ten-bin expected calibration error of the top probability; pooled over all 3,880 questions"
 },
 "note": "Ollaya Bench edition 2026.10.",
 "results": [
  {
   "server": "ollaya",
   "model": "decider:4b",
   "digest": "b4e1276b55eda06fad7af9536a996fc132a1229bb866755cd6f07778dbe98963",
   "records": 3880,
   "errors": 0,
   "macro_accuracy": 0.769,
   "macro_accuracy_ci": [
    0.7561,
    0.7812
   ],
   "micro_accuracy": 0.7807,
   "macro_ece": 0.1375,
   "pooled_ece": 0.0944,
   "brier": 0.3532,
   "median_ms": 110.4,
   "p95_ms": 655.1,
   "by_type": {
    "choice": 0.8349,
    "noul": 0.8237,
    "score": 0.5679
   },
   "per_subset": {
    "vitaminc-dev": {
     "n": 599,
     "errors": 0,
     "accuracy": 0.7913,
     "ece": 0.0802
    },
    "massive-en-US": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.8657,
     "ece": 0.0556
    },
    "massive-de-DE": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.86,
     "ece": 0.0659
    },
    "boolq": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.86,
     "ece": 0.139
    },
    "squad2": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.8094,
     "ece": 0.1873
    },
    "paws": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.792,
     "ece": 0.2112
    },
    "multinli": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.9097,
     "ece": 0.0372
    },
    "civil_comments": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.8533,
     "ece": 0.146
    },
    "aegis2": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.804,
     "ece": 0.1992
    },
    "helpsteer2": {
     "n": 249,
     "errors": 0,
     "accuracy": 0.4538,
     "ece": 0.0571
    },
    "summeval-relevance": {
     "n": 240,
     "errors": 0,
     "accuracy": 0.4583,
     "ece": 0.1182
    },
    "summeval-consistency": {
     "n": 144,
     "errors": 0,
     "accuracy": 0.7917,
     "ece": 0.4237
    },
    "pubmedqa": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.748,
     "ece": 0.0672
    }
   }
  },
  {
   "server": "ollaya",
   "model": "strands-decider:e4b",
   "digest": "b923eec149f21f2b2cc2a294c9b1da22c52cbbdb72b0f6d044e63054765809c3",
   "records": 3880,
   "errors": 0,
   "macro_accuracy": 0.7648,
   "macro_accuracy_ci": [
    0.752,
    0.7779
   ],
   "micro_accuracy": 0.7719,
   "macro_ece": 0.0925,
   "pooled_ece": 0.0392,
   "brier": 0.3245,
   "median_ms": 115.1,
   "p95_ms": 205.6,
   "by_type": {
    "choice": 0.8049,
    "noul": 0.8463,
    "score": 0.5623
   },
   "per_subset": {
    "vitaminc-dev": {
     "n": 599,
     "errors": 0,
     "accuracy": 0.7663,
     "ece": 0.046
    },
    "massive-en-US": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.8657,
     "ece": 0.0431
    },
    "massive-de-DE": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.84,
     "ece": 0.0394
    },
    "boolq": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.87,
     "ece": 0.038
    },
    "squad2": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.8094,
     "ece": 0.03
    },
    "paws": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.88,
     "ece": 0.031
    },
    "multinli": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.8127,
     "ece": 0.1208
    },
    "civil_comments": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.92,
     "ece": 0.0296
    },
    "aegis2": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.752,
     "ece": 0.1079
    },
    "helpsteer2": {
     "n": 249,
     "errors": 0,
     "accuracy": 0.3494,
     "ece": 0.0452
    },
    "summeval-relevance": {
     "n": 240,
     "errors": 0,
     "accuracy": 0.4625,
     "ece": 0.1668
    },
    "summeval-consistency": {
     "n": 144,
     "errors": 0,
     "accuracy": 0.875,
     "ece": 0.4354
    },
    "pubmedqa": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.74,
     "ece": 0.0699
    }
   }
  },
  {
   "server": "ollaya",
   "model": "gliclass-v4:9b",
   "digest": "0e5bbb07b002e136f66bd57e8dd5bf42613df06dfbb55c80ebd016019e737529",
   "records": 3880,
   "errors": 0,
   "macro_accuracy": 0.7628,
   "macro_accuracy_ci": [
    0.7505,
    0.7747
   ],
   "micro_accuracy": 0.7688,
   "macro_ece": 0.0731,
   "pooled_ece": 0.0338,
   "brier": 0.3148,
   "median_ms": 158.8,
   "p95_ms": 301.1,
   "by_type": {
    "choice": 0.8212,
    "noul": 0.8618,
    "score": 0.5003
   },
   "per_subset": {
    "vitaminc-dev": {
     "n": 599,
     "errors": 0,
     "accuracy": 0.7129,
     "ece": 0.1233
    },
    "massive-en-US": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.8771,
     "ece": 0.0251
    },
    "massive-de-DE": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.8571,
     "ece": 0.0284
    },
    "boolq": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.86,
     "ece": 0.0442
    },
    "squad2": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.8562,
     "ece": 0.0355
    },
    "paws": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.868,
     "ece": 0.0218
    },
    "multinli": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.8829,
     "ece": 0.0795
    },
    "civil_comments": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.9167,
     "ece": 0.0262
    },
    "aegis2": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.808,
     "ece": 0.0573
    },
    "helpsteer2": {
     "n": 249,
     "errors": 0,
     "accuracy": 0.3855,
     "ece": 0.1334
    },
    "summeval-relevance": {
     "n": 240,
     "errors": 0,
     "accuracy": 0.275,
     "ece": 0.1979
    },
    "summeval-consistency": {
     "n": 144,
     "errors": 0,
     "accuracy": 0.8403,
     "ece": 0.142
    },
    "pubmedqa": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.776,
     "ece": 0.0362
    }
   }
  },
  {
   "server": "ollaya",
   "model": "gliclass-v4:2b",
   "digest": "e2575bb26e1362f95d78049e2988ba36973ebb95f308b7d3c94aa277dd05a870",
   "records": 3880,
   "errors": 0,
   "macro_accuracy": 0.7309,
   "macro_accuracy_ci": [
    0.7182,
    0.7439
   ],
   "micro_accuracy": 0.7379,
   "macro_ece": 0.0886,
   "pooled_ece": 0.0307,
   "brier": 0.3598,
   "median_ms": 76.9,
   "p95_ms": 105.3,
   "by_type": {
    "choice": 0.765,
    "noul": 0.8332,
    "score": 0.5036
   },
   "per_subset": {
    "vitaminc-dev": {
     "n": 599,
     "errors": 0,
     "accuracy": 0.7162,
     "ece": 0.0822
    },
    "massive-en-US": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.8257,
     "ece": 0.058
    },
    "massive-de-DE": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.7943,
     "ece": 0.0503
    },
    "boolq": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.8033,
     "ece": 0.075
    },
    "squad2": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.8462,
     "ece": 0.018
    },
    "paws": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.816,
     "ece": 0.0582
    },
    "multinli": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.8127,
     "ece": 0.0835
    },
    "civil_comments": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.8967,
     "ece": 0.0157
    },
    "aegis2": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.804,
     "ece": 0.108
    },
    "helpsteer2": {
     "n": 249,
     "errors": 0,
     "accuracy": 0.4177,
     "ece": 0.1365
    },
    "summeval-relevance": {
     "n": 240,
     "errors": 0,
     "accuracy": 0.2667,
     "ece": 0.1618
    },
    "summeval-consistency": {
     "n": 144,
     "errors": 0,
     "accuracy": 0.8264,
     "ece": 0.2579
    },
    "pubmedqa": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.676,
     "ece": 0.0468
    }
   }
  },
  {
   "server": "ollaya",
   "model": "decider:2b",
   "digest": "6b81ac65a337f2888699e594e1f325aefe7e9718ce0fba8d6da4a3492aba1982",
   "records": 3880,
   "errors": 0,
   "macro_accuracy": 0.7235,
   "macro_accuracy_ci": [
    0.7096,
    0.7372
   ],
   "micro_accuracy": 0.7317,
   "macro_ece": 0.1511,
   "pooled_ece": 0.0888,
   "brier": 0.4115,
   "median_ms": 69.9,
   "p95_ms": 217.7,
   "by_type": {
    "choice": 0.7848,
    "noul": 0.7925,
    "score": 0.5062
   },
   "per_subset": {
    "vitaminc-dev": {
     "n": 599,
     "errors": 0,
     "accuracy": 0.7045,
     "ece": 0.0646
    },
    "massive-en-US": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.82,
     "ece": 0.0739
    },
    "massive-de-DE": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.8171,
     "ece": 0.1138
    },
    "boolq": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.83,
     "ece": 0.1692
    },
    "squad2": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.7726,
     "ece": 0.2276
    },
    "paws": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.72,
     "ece": 0.2774
    },
    "multinli": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.8662,
     "ece": 0.0392
    },
    "civil_comments": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.9,
     "ece": 0.0989
    },
    "aegis2": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.74,
     "ece": 0.2608
    },
    "helpsteer2": {
     "n": 249,
     "errors": 0,
     "accuracy": 0.4337,
     "ece": 0.0741
    },
    "summeval-relevance": {
     "n": 240,
     "errors": 0,
     "accuracy": 0.2792,
     "ece": 0.0717
    },
    "summeval-consistency": {
     "n": 144,
     "errors": 0,
     "accuracy": 0.8056,
     "ece": 0.4074
    },
    "pubmedqa": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.716,
     "ece": 0.0858
    }
   }
  },
  {
   "server": "ollaya",
   "model": "strands-decider:e2b",
   "digest": "76e87da292235d224abe1f24f19596c5280c9d9eabe4e0ff3278697de088d438",
   "records": 3880,
   "errors": 0,
   "macro_accuracy": 0.7161,
   "macro_accuracy_ci": [
    0.7015,
    0.7284
   ],
   "micro_accuracy": 0.7227,
   "macro_ece": 0.131,
   "pooled_ece": 0.0653,
   "brier": 0.394,
   "median_ms": 78.2,
   "p95_ms": 114.7,
   "by_type": {
    "choice": 0.7238,
    "noul": 0.8372,
    "score": 0.5016
   },
   "per_subset": {
    "vitaminc-dev": {
     "n": 599,
     "errors": 0,
     "accuracy": 0.7229,
     "ece": 0.0653
    },
    "massive-en-US": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.8314,
     "ece": 0.0841
    },
    "massive-de-DE": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.7057,
     "ece": 0.1851
    },
    "boolq": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.8033,
     "ece": 0.0578
    },
    "squad2": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.7592,
     "ece": 0.043
    },
    "paws": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.932,
     "ece": 0.2116
    },
    "multinli": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.699,
     "ece": 0.04
    },
    "civil_comments": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.9033,
     "ece": 0.1351
    },
    "aegis2": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.788,
     "ece": 0.0553
    },
    "helpsteer2": {
     "n": 249,
     "errors": 0,
     "accuracy": 0.3133,
     "ece": 0.0266
    },
    "summeval-relevance": {
     "n": 240,
     "errors": 0,
     "accuracy": 0.4208,
     "ece": 0.1889
    },
    "summeval-consistency": {
     "n": 144,
     "errors": 0,
     "accuracy": 0.7708,
     "ece": 0.4985
    },
    "pubmedqa": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.66,
     "ece": 0.1121
    }
   }
  },
  {
   "server": "ollaya",
   "model": "d1-omni:600m",
   "digest": "53dfa4be304b8ab233617f432ed655050d6d459a48f89dae94a574f75414f139",
   "records": 3880,
   "errors": 0,
   "macro_accuracy": 0.6393,
   "macro_accuracy_ci": [
    0.6251,
    0.6535
   ],
   "micro_accuracy": 0.6644,
   "macro_ece": 0.1071,
   "pooled_ece": 0.0748,
   "brier": 0.4464,
   "median_ms": 14.8,
   "p95_ms": 36,
   "by_type": {
    "choice": 0.7314,
    "noul": 0.732,
    "score": 0.3313
   },
   "per_subset": {
    "vitaminc-dev": {
     "n": 599,
     "errors": 0,
     "accuracy": 0.6594,
     "ece": 0.0751
    },
    "massive-en-US": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.8371,
     "ece": 0.0842
    },
    "massive-de-DE": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.7171,
     "ece": 0.0865
    },
    "boolq": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.69,
     "ece": 0.1991
    },
    "squad2": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.7659,
     "ece": 0.1086
    },
    "paws": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.704,
     "ece": 0.048
    },
    "multinli": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.7993,
     "ece": 0.0912
    },
    "civil_comments": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.86,
     "ece": 0.1086
    },
    "aegis2": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.64,
     "ece": 0.1635
    },
    "helpsteer2": {
     "n": 249,
     "errors": 0,
     "accuracy": 0.4217,
     "ece": 0.1005
    },
    "summeval-relevance": {
     "n": 240,
     "errors": 0,
     "accuracy": 0.225,
     "ece": 0.0412
    },
    "summeval-consistency": {
     "n": 144,
     "errors": 0,
     "accuracy": 0.3472,
     "ece": 0.1281
    },
    "pubmedqa": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.644,
     "ece": 0.1571
    }
   }
  },
  {
   "server": "ollaya",
   "model": "sura:0.8b",
   "digest": "473fa350466df82396f2a19152bdc15abaa79a6e57fc99f867c3e0dd280ae18a",
   "records": 3880,
   "errors": 0,
   "macro_accuracy": 0.5861,
   "macro_accuracy_ci": [
    0.5716,
    0.6012
   ],
   "micro_accuracy": 0.6088,
   "macro_ece": 0.1324,
   "pooled_ece": 0.0573,
   "brier": 0.509,
   "median_ms": 72.7,
   "p95_ms": 91.2,
   "by_type": {
    "choice": 0.6396,
    "noul": 0.6672,
    "score": 0.3619
   },
   "per_subset": {
    "vitaminc-dev": {
     "n": 599,
     "errors": 0,
     "accuracy": 0.7212,
     "ece": 0.0382
    },
    "massive-en-US": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.7514,
     "ece": 0.0299
    },
    "massive-de-DE": {
     "n": 350,
     "errors": 0,
     "accuracy": 0.6171,
     "ece": 0.066
    },
    "boolq": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.6967,
     "ece": 0.0495
    },
    "squad2": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.6087,
     "ece": 0.214
    },
    "paws": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.664,
     "ece": 0.0657
    },
    "multinli": {
     "n": 299,
     "errors": 0,
     "accuracy": 0.6321,
     "ece": 0.071
    },
    "civil_comments": {
     "n": 300,
     "errors": 0,
     "accuracy": 0.8467,
     "ece": 0.0449
    },
    "aegis2": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.52,
     "ece": 0.2325
    },
    "helpsteer2": {
     "n": 249,
     "errors": 0,
     "accuracy": 0.3052,
     "ece": 0.158
    },
    "summeval-relevance": {
     "n": 240,
     "errors": 0,
     "accuracy": 0.1417,
     "ece": 0.4137
    },
    "summeval-consistency": {
     "n": 144,
     "errors": 0,
     "accuracy": 0.6389,
     "ece": 0.262
    },
    "pubmedqa": {
     "n": 250,
     "errors": 0,
     "accuracy": 0.476,
     "ece": 0.0753
    }
   }
  }
 ]
}
