{
  "scope": "Solver inference only. The model ran on a Claude Max subscription through Claude Code, so these are list-price estimates, not bills: Claude Code's own list-price total for each session, the basis of every Claude column on the board. Judging is excluded. Every run in the sweep is priced.",
  "pricing_method": "Claude Code's own list-price total for the session (the last result receipt's total_cost_usd, which is cumulative over the session's receipts), the cost basis of every Claude column; the adapter's own estimate prices cached prompt tokens at the full input rate and is not used",
  "groups": [
    {
      "model": "fable52",
      "effort": "low",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 1.723948452173913,
        "se": 0.2093140689825835
      },
      "usd_total": 39.65081439999999,
      "raw_tokens": {
        "n": 23,
        "mean": 648534.1739130435,
        "se": 151670.6535935045
      },
      "raw_tokens_total": 14916286,
      "output_tokens": {
        "n": 23,
        "mean": 16184.173913043478,
        "se": 2162.1580491386917
      },
      "thinking_tokens": {
        "n": 23,
        "mean": 7713.521739130435,
        "se": 1115.0474154110589
      },
      "minutes": {
        "n": 23,
        "mean": 4.633498550724638,
        "se": 0.6441842109354138
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "fable52",
      "effort": "medium",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 1.717883243478261,
        "se": 0.18482904926442256
      },
      "usd_total": 39.51131459999999,
      "raw_tokens": {
        "n": 23,
        "mean": 593073.5652173914,
        "se": 99999.30673221798
      },
      "raw_tokens_total": 13640692,
      "output_tokens": {
        "n": 23,
        "mean": 15047.130434782608,
        "se": 1100.128438970025
      },
      "thinking_tokens": {
        "n": 23,
        "mean": 7081.086956521739,
        "se": 633.5495616659588
      },
      "minutes": {
        "n": 23,
        "mean": 4.298171014492754,
        "se": 0.45573022195522306
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "fable52",
      "effort": "high",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 2.8564122086956525,
        "se": 0.2750240187769356
      },
      "usd_total": 65.69748080000001,
      "raw_tokens": {
        "n": 23,
        "mean": 1188268.2608695652,
        "se": 197518.33191204575
      },
      "raw_tokens_total": 27330170,
      "output_tokens": {
        "n": 23,
        "mean": 30809.130434782608,
        "se": 2108.2419265743206
      },
      "thinking_tokens": {
        "n": 23,
        "mean": 18275.478260869564,
        "se": 1344.2933313638323
      },
      "minutes": {
        "n": 23,
        "mean": 7.8015108695652176,
        "se": 0.7607266948369847
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "fable52",
      "effort": "extra-high",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 4.989511504347826,
        "se": 0.3580790396223007
      },
      "usd_total": 114.75876459999999,
      "raw_tokens": {
        "n": 23,
        "mean": 2510352.347826087,
        "se": 334833.8805782199
      },
      "raw_tokens_total": 57738104,
      "output_tokens": {
        "n": 23,
        "mean": 61658.913043478264,
        "se": 3245.1431796722463
      },
      "thinking_tokens": {
        "n": 23,
        "mean": 40767.782608695656,
        "se": 2258.4977263803708
      },
      "minutes": {
        "n": 23,
        "mean": 14.200776811594203,
        "se": 0.752536884888609
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "fable52",
      "effort": "max",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 10.8796342,
        "se": 0.43585059797352127
      },
      "usd_total": 250.23158659999999,
      "raw_tokens": {
        "n": 23,
        "mean": 7905905.565217392,
        "se": 557017.021264683
      },
      "raw_tokens_total": 181835828,
      "output_tokens": {
        "n": 23,
        "mean": 141517.26086956522,
        "se": 5125.203639342809
      },
      "thinking_tokens": {
        "n": 23,
        "mean": 94675.52173913043,
        "se": 3583.3733587142365
      },
      "minutes": {
        "n": 23,
        "mean": 37.08653188405797,
        "se": 2.1561391816650817
      },
      "solver_fallback_runs": 0
    }
  ],
  "totals": {
    "fable52": {
      "runs": 115,
      "usd": 509.8499610000001,
      "raw_tokens": 295461080,
      "solver_hours": 26.074520833333317
    }
  },
  "limitations": [
    "Claude Code computes the total at its own list rates from the session's usage and reports it cumulatively on each result receipt; the export takes the session's last total (two runs have two receipts) and matched every run to the population record.",
    "The adapter's own run-time estimate prices cached prompt tokens at the full input rate and reads higher (Max $14.01 a task against $10.88); it is not used.",
    "The estimate covers the solver's exposed receipts, not an independently observed invoice."
  ],
  "note": "Per-run estimates are in runs.json (estimated_usd, raw_tokens, token_usage). Judging is excluded."
}