{
  "metadata": {
    "benchmark": "RollerCoasterTycoonBench",
    "slug": "rctbench",
    "description": "Can a decision model run a theme park? Nine models manage a RollerCoaster Tycoon 2 park for three in-game years.",
    "benchmark_id": "rctbench",
    "family": "rctbench",
    "version": "1",
    "updated": "2026-11-19",
    "dataset_type": "private",
    "industry": "beta",
    "total_models": 9,
    "models": [
      "decision/decider-flash-v1.2",
      "decision/decider-flash-v1.3",
      "decision/decider-v1.2",
      "decision/decider-v1.3",
      "decision/drex-max",
      "decision/gemini-decision-1",
      "decision/gpt-6-luna",
      "decision/jev-1.13.0",
      "decision/jev-1.14"
    ],
    "visible": true,
    "showBadge": false,
    "show_stderr": true,
    "latency_label": "Time / decision",
    "partners": [],
    "use_cost_per_test": true,
    "cost_label": "Cost / run",
    "runner": "external",
    "mode": "agentic",
    "log_scale": true,
    "archived": false,
    "partner": false,
    "tasks": {
      "overall": "Overall"
    }
  },
  "tasks": {
    "overall": {
      "decision/decider-v1.3": {
        "accuracy": 70.963,
        "latency": 2.543,
        "stderr": 2.738,
        "cost_per_test": 0.2925,
        "temperature": null,
        "top_p": null,
        "max_output_tokens": null,
        "reasoning": false,
        "reasoning_effort": null,
        "verbosity": null,
        "compute_effort": null,
        "provider": "Perplexity",
        "harness": "rctbench"
      },
      "decision/decider-v1.2": {
        "accuracy": 67.06,
        "latency": 1.3,
        "stderr": 3.403,
        "cost_per_test": 0.211867,
        "temperature": null,
        "top_p": null,
        "max_output_tokens": null,
        "reasoning": false,
        "reasoning_effort": null,
        "verbosity": null,
        "compute_effort": null,
        "provider": "Perplexity",
        "harness": "rctbench"
      },
      "decision/jev-1.13.0": {
        "accuracy": 59.223,
        "latency": 0.565,
        "stderr": 6.728,
        "cost_per_test": 2.1833,
        "temperature": null,
        "top_p": null,
        "max_output_tokens": null,
        "reasoning": false,
        "reasoning_effort": null,
        "verbosity": null,
        "compute_effort": null,
        "provider": "TypeSafe",
        "harness": "rctbench"
      },
      "decision/drex-max": {
        "accuracy": 56.87,
        "latency": 1.263,
        "stderr": 4.714,
        "cost_per_test": 0.369067,
        "temperature": null,
        "top_p": null,
        "max_output_tokens": null,
        "reasoning": false,
        "reasoning_effort": null,
        "verbosity": null,
        "compute_effort": null,
        "provider": "Nace.AI",
        "harness": "rctbench"
      },
      "decision/gpt-6-luna": {
        "accuracy": 55.443,
        "latency": 0.591,
        "stderr": 10.108,
        "cost_per_test": 3.758933,
        "temperature": null,
        "top_p": null,
        "max_output_tokens": null,
        "reasoning": false,
        "reasoning_effort": null,
        "verbosity": null,
        "compute_effort": null,
        "provider": "OpenAI",
        "harness": "rctbench"
      },
      "decision/jev-1.14": {
        "accuracy": 42.743,
        "latency": 0.524,
        "stderr": 10.782,
        "cost_per_test": 1.070833,
        "temperature": null,
        "top_p": null,
        "max_output_tokens": null,
        "reasoning": false,
        "reasoning_effort": null,
        "verbosity": null,
        "compute_effort": null,
        "provider": "TypeSafe",
        "harness": "rctbench"
      },
      "decision/gemini-decision-1": {
        "accuracy": 29.13,
        "latency": 1.126,
        "stderr": 6.323,
        "cost_per_test": 1.124567,
        "temperature": null,
        "top_p": null,
        "max_output_tokens": null,
        "reasoning": false,
        "reasoning_effort": null,
        "verbosity": null,
        "compute_effort": null,
        "provider": "Google",
        "harness": "rctbench"
      },
      "decision/decider-flash-v1.3": {
        "accuracy": 21.427,
        "latency": 0.586,
        "stderr": 10.055,
        "cost_per_test": 0.1029,
        "temperature": null,
        "top_p": null,
        "max_output_tokens": null,
        "reasoning": false,
        "reasoning_effort": null,
        "verbosity": null,
        "compute_effort": null,
        "provider": "Perplexity",
        "harness": "rctbench"
      },
      "decision/decider-flash-v1.2": {
        "accuracy": 5.31,
        "latency": 6.393,
        "stderr": 2.192,
        "cost_per_test": 0.530633,
        "temperature": null,
        "top_p": null,
        "max_output_tokens": null,
        "reasoning": false,
        "reasoning_effort": null,
        "verbosity": null,
        "compute_effort": null,
        "provider": "Perplexity",
        "harness": "rctbench"
      }
    }
  }
}
