{
  "benchmark": "HuDetect-1k",
  "version": "1.3",
  "measured": "2026-09",
  "dataset": "https://huggingface.co/datasets/mihalykiss/GenAIDetectHU",
  "human_texts": 500,
  "ai_texts": 500,
  "positive_class": "AI-generated",
  "disclosure": "Dataset and benchmark created by the developer of Preds, one of the compared detectors.",
  "license": "CC BY 4.0",
  "results": [
    {
      "detector": "Preds 3",
      "accuracy": 99.6,
      "f1": 99.6,
      "auroc": 0.9996,
      "fpr": 0.0,
      "fnr": 0.8,
      "precision": 100.0,
      "recall": 99.2,
      "tp": 496,
      "fp": 0,
      "tn": 500,
      "fn": 4
    },
    {
      "detector": "GPTZero",
      "accuracy": 90.1,
      "f1": 89.04,
      "auroc": 0.9923,
      "fpr": 0.2,
      "fnr": 19.6,
      "precision": 99.75,
      "recall": 80.4,
      "tp": 402,
      "fp": 1,
      "tn": 499,
      "fn": 98
    },
    {
      "detector": "Originality.ai",
      "accuracy": 82.0,
      "f1": 84.29,
      "auroc": 0.9574,
      "fpr": 32.6,
      "fnr": 3.4,
      "precision": 74.77,
      "recall": 96.6,
      "tp": 483,
      "fp": 163,
      "tn": 337,
      "fn": 17
    },
    {
      "detector": "Plag.hu",
      "accuracy": 76.5,
      "f1": 69.44,
      "auroc": 0.948,
      "fpr": 0.4,
      "fnr": 46.6,
      "precision": 99.26,
      "recall": 53.4,
      "tp": 267,
      "fp": 2,
      "tn": 498,
      "fn": 233
    },
    {
      "detector": "isgen.ai",
      "accuracy": 65.5,
      "f1": 48.43,
      "auroc": 0.7317,
      "fpr": 1.4,
      "fnr": 67.6,
      "precision": 95.86,
      "recall": 32.4,
      "tp": 162,
      "fp": 7,
      "tn": 493,
      "fn": 338
    },
    {
      "detector": "ZeroGPT",
      "accuracy": 64.7,
      "f1": 64.52,
      "auroc": 0.6911,
      "fpr": 34.8,
      "fnr": 35.8,
      "precision": 64.85,
      "recall": 64.2,
      "tp": 321,
      "fp": 174,
      "tn": 326,
      "fn": 179
    },
    {
      "detector": "Smodin",
      "accuracy": 62.9,
      "f1": 41.39,
      "auroc": 0.7655,
      "fpr": 0.4,
      "fnr": 73.8,
      "precision": 98.5,
      "recall": 26.2,
      "tp": 131,
      "fp": 2,
      "tn": 498,
      "fn": 369
    }
  ]
}