{
  "scope": "Solver inference only, API-equivalent at list rates; both models ran on a ChatGPT subscription through Codex, so these are estimates, not bills. Judging and local infrastructure are excluded.",
  "pricing_verified": "2026-09-11",
  "sources": {
    "openai": "https://developers.openai.com/api/docs/pricing"
  },
  "rates_per_million": {
    "gpt55": {
      "input": 5.0,
      "cached_input": 0.5,
      "output": 30.0
    },
    "luna": {
      "input": 0.2,
      "cached_input": 0.02,
      "output": 1.2
    }
  },
  "groups": [
    {
      "model": "gpt55",
      "effort": "low",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 1.7947075652173912,
        "se": 0.1926970452424152
      },
      "usd_total": 41.278273999999996,
      "raw_tokens": {
        "n": 23,
        "mean": 1715428.347826087,
        "se": 232298.7180522651
      },
      "raw_tokens_total": 39454852,
      "minutes": {
        "n": 23,
        "mean": 8.210286231884059,
        "se": 0.8665636722344808
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "gpt55",
      "effort": "medium",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 2.708066043478261,
        "se": 0.5016670652294555
      },
      "usd_total": 62.285519,
      "raw_tokens": {
        "n": 23,
        "mean": 2740785.652173913,
        "se": 635095.1076762121
      },
      "raw_tokens_total": 63038070,
      "minutes": {
        "n": 23,
        "mean": 12.403088405797101,
        "se": 1.9385437189805725
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "gpt55",
      "effort": "high",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 4.197308130434783,
        "se": 0.764545759697286
      },
      "usd_total": 96.538087,
      "raw_tokens": {
        "n": 23,
        "mean": 4516861,
        "se": 1005272.6914506525
      },
      "raw_tokens_total": 103887803,
      "minutes": {
        "n": 23,
        "mean": 18.245584782608695,
        "se": 2.4628805564760197
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "gpt55",
      "effort": "extra-high",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 4.562569521739131,
        "se": 0.5698215364584018
      },
      "usd_total": 104.939099,
      "raw_tokens": {
        "n": 23,
        "mean": 4718556.173913044,
        "se": 744132.3102685512
      },
      "raw_tokens_total": 108526792,
      "minutes": {
        "n": 23,
        "mean": 20.783839855072465,
        "se": 2.067568406636599
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "luna",
      "effort": "low",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 0.023995782608695653,
        "se": 0.0012330916060532804
      },
      "usd_total": 0.551903,
      "raw_tokens": {
        "n": 23,
        "mean": 422906.2173913043,
        "se": 28475.850213735237
      },
      "raw_tokens_total": 9726843,
      "minutes": {
        "n": 23,
        "mean": 2.662527536231884,
        "se": 0.10900985778255687
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "luna",
      "effort": "medium",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 0.06639808695652175,
        "se": 0.00437247942519653
      },
      "usd_total": 1.527156,
      "raw_tokens": {
        "n": 23,
        "mean": 1347575.1739130435,
        "se": 119662.27067229344
      },
      "raw_tokens_total": 30994229,
      "minutes": {
        "n": 23,
        "mean": 7.397060144927536,
        "se": 0.40243248052154174
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "luna",
      "effort": "high",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 0.21676852173913044,
        "se": 0.029096109421329557
      },
      "usd_total": 4.985676,
      "raw_tokens": {
        "n": 23,
        "mean": 5889062.130434782,
        "se": 990551.6134088954
      },
      "raw_tokens_total": 135448429,
      "minutes": {
        "n": 23,
        "mean": 20.99510579710145,
        "se": 2.0721589235938365
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "luna",
      "effort": "extra-high",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 0.2747059565217391,
        "se": 0.03819344758449075
      },
      "usd_total": 6.318237,
      "raw_tokens": {
        "n": 23,
        "mean": 7581839,
        "se": 1337409.5792276198
      },
      "raw_tokens_total": 174382297,
      "minutes": {
        "n": 23,
        "mean": 25.42935579710145,
        "se": 2.6095592085448867
      },
      "solver_fallback_runs": 0
    },
    {
      "model": "luna",
      "effort": "max",
      "n": 23,
      "usd": {
        "n": 23,
        "mean": 0.448386,
        "se": 0.12004178943637486
      },
      "usd_total": 10.312878,
      "raw_tokens": {
        "n": 23,
        "mean": 11882951.652173912,
        "se": 3482159.2859923877
      },
      "raw_tokens_total": 273307888,
      "minutes": {
        "n": 23,
        "mean": 44.09918623188406,
        "se": 10.719661343012751
      },
      "solver_fallback_runs": 0
    }
  ],
  "totals": {
    "gpt55": {
      "runs": 92,
      "usd": 305.040979,
      "raw_tokens": 314907517,
      "solver_hours": 22.863073055555553
    },
    "luna": {
      "runs": 115,
      "usd": 23.69585,
      "raw_tokens": 623859686,
      "solver_hours": 38.55690694444444
    }
  },
  "limitations": [
    "Codex receipts report input, cached input, output and reasoning tokens per run; cached input is billed at the cache-read rate and the rest of the input at the standard rate. Cache writes are free on this API and are not modelled.",
    "Standard tier, short-context rates. Per-request context sizes are not exposed by the receipts, so no long-context premium is applied.",
    "No Batch, Flex, Fast or priority pricing is applied.",
    "The estimate covers the solver's exposed receipts, not an independently observed API invoice.",
    "One GPT-5.5 extra-high run (paddockcore) was re-run after the first attempt overran its cap under a harness fault; the re-run is the priced and judged run and the capped attempt is excluded."
  ],
  "comparison_sha256": "cc114b736cbdad27595c1a9154ab042acfa3239f373726aabe59f88fa0531b38",
  "note": "Per-run estimates are in runs.json (estimated_usd, raw_tokens, token_usage). Judging is excluded."
}