{
  "schema": "fresh402-agent-efficiency-v1",
  "recorded_date": "2026-11-05",
  "task": "Controlled65-operation OpenAPI3.1 review; removed GET operation, cursor parameter becomes required, parameter ordering reversed.",
  "model": "gpt-6.1-sol",
  "paired_trials": 5,
  "pricing": {
    "source": "https://developers.openai.com/api/docs/models/gpt-6.1-sol",
    "retrieved": "2026-11-05",
    "input_per_million": 2,
    "cached_input_per_million": 0.1,
    "output_per_million": 10,
    "cache_write_per_million": 2.5,
    "basis": "Standard API equivalent, not an actual ChatGPT/Codex subscription invoice. No billed API invoked.",
    "prepaid": "Starter2.50/1000 and Developer15/10000 consumed-credit equivalents. Upfront unused stored value excluded from marginal task fees, not silently consumed."
  },
  "median_summary": {
    "free": {
      "n": 5,
      "passed": 5,
      "input_tokens": {
        "min": 42253,
        "median": 49578,
        "max": 49942
      },
      "cached_input_tokens": {
        "min": 23296,
        "median": 34432,
        "max": 34560
      },
      "output_tokens": {
        "min": 510,
        "median": 540,
        "max": 566
      },
      "total_tokens": {
        "min": 42819,
        "median": 50113,
        "max": 50482
      },
      "wall_ms": {
        "min": 20541,
        "median": 21502,
        "max": 24328
      },
      "tool_calls": {
        "min": 4,
        "median": 4,
        "max": 4
      },
      "inference_api_equivalent_usd": {
        "min": 0.0313116,
        "median": 0.0390852,
        "max": 0.0460236
      },
      "starter_total_equivalent_usd": {
        "min": 0.0313116,
        "median": 0.0390852,
        "max": 0.0460236
      },
      "developer_total_equivalent_usd": {
        "min": 0.0313116,
        "median": 0.0390852,
        "max": 0.0460236
      },
      "direct_total_equivalent_usd": {
        "min": 0.0313116,
        "median": 0.0390852,
        "max": 0.0460236
      },
      "completed_sample_only": true
    },
    "fresh402": {
      "n": 5,
      "passed": 5,
      "input_tokens": {
        "min": 41224,
        "median": 42021,
        "max": 49984
      },
      "cached_input_tokens": {
        "min": 20096,
        "median": 30336,
        "max": 30976
      },
      "output_tokens": {
        "min": 516,
        "median": 555,
        "max": 605
      },
      "total_tokens": {
        "min": 41829,
        "median": 42537,
        "max": 50565
      },
      "wall_ms": {
        "min": 20299,
        "median": 21887,
        "max": 23343
      },
      "tool_calls": {
        "min": 3,
        "median": 4,
        "max": 4
      },
      "inference_api_equivalent_usd": {
        "min": 0.030859599999999997,
        "median": 0.0314956,
        "max": 0.0527604
      },
      "starter_total_equivalent_usd": {
        "min": 0.0433596,
        "median": 0.043995599999999996,
        "max": 0.0652604
      },
      "developer_total_equivalent_usd": {
        "min": 0.038359599999999994,
        "median": 0.0389956,
        "max": 0.0602604
      },
      "direct_total_equivalent_usd": {
        "min": 0.0508596,
        "median": 0.0514956,
        "max": 0.0727604
      },
      "completed_sample_only": true
    },
    "comparison": {
      "token_reduction_fraction": 0.15117833695847382,
      "latency_reduction_fraction": -0.01790531113384808,
      "starter_cost_reduction_fraction": -0.12563323201621057,
      "developer_cost_reduction_fraction": 0.002292427824342802,
      "cost_win": false,
      "break_even": "No guaranteed break-even where the competent baseline also reduces model-visible context; compute only conditional on a positive measured inference delta.",
      "projections_5_10_20": "Not published: recurring-cycle costs and production network/cold starts were not measured by this controlled fixture comparison."
    }
  },
  "methodology": "Fresh independent LLM contexts chose tools and interpreted evidence. Same synthetic source contracts, semantic correctness requirements and model. Free baseline: installed oasdiff1.34.1 breaking+diff. Fresh402: current production comparison functions in local workerd, not production network, settlement or Container latency. Model tool-description context, baseline parsing and local process startup included; pre-existing dependency installation excluded equally. Cached inputs priced separately, reasoning already included in output. Service prices/credit costs are equivalents only: no real local fee/debit. Upfront unused prepaid stored value remains unspent. Complete10-run API sample retained. No claim about unrelated workflows, general superiority or realized cash savings.",
  "runs": [
    {
      "task": "api-1",
      "condition": "free",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 50113,
        "inputTokens": 49578,
        "cachedInputTokens": 34432,
        "cacheWriteInputTokens": 0,
        "outputTokens": 535,
        "reasoningOutputTokens": 13,
        "noncachedInputTokens": 15146,
        "reasoning_counted_in_output": true
      },
      "model_requests": 5,
      "tool_calls": 4,
      "retries": 0,
      "wall_ms": 20736,
      "inference_api_equivalent_usd": 0.0390852,
      "direct_fee_equivalent_usd": 0,
      "prepaid_credits_equivalent": 0,
      "starter_total_equivalent_usd": 0.0390852,
      "developer_total_equivalent_usd": 0.0390852,
      "direct_total_equivalent_usd": 0.0390852,
      "actual_credit_debits": 0
    },
    {
      "task": "api-1",
      "condition": "fresh402",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 42955,
        "inputTokens": 42400,
        "cachedInputTokens": 30976,
        "cacheWriteInputTokens": 0,
        "outputTokens": 555,
        "reasoningOutputTokens": 25,
        "noncachedInputTokens": 11424,
        "reasoning_counted_in_output": true
      },
      "model_requests": 5,
      "tool_calls": 4,
      "retries": 0,
      "wall_ms": 21887,
      "inference_api_equivalent_usd": 0.0314956,
      "direct_fee_equivalent_usd": 0.02,
      "prepaid_credits_equivalent": 5,
      "starter_total_equivalent_usd": 0.043995599999999996,
      "developer_total_equivalent_usd": 0.0389956,
      "direct_total_equivalent_usd": 0.0514956,
      "actual_credit_debits": 0
    },
    {
      "task": "api-2",
      "condition": "free",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 50287,
        "inputTokens": 49777,
        "cachedInputTokens": 34560,
        "cacheWriteInputTokens": 0,
        "outputTokens": 510,
        "reasoningOutputTokens": 14,
        "noncachedInputTokens": 15217,
        "reasoning_counted_in_output": true
      },
      "model_requests": 5,
      "tool_calls": 4,
      "retries": 0,
      "wall_ms": 20541,
      "inference_api_equivalent_usd": 0.03899,
      "direct_fee_equivalent_usd": 0,
      "prepaid_credits_equivalent": 0,
      "starter_total_equivalent_usd": 0.03899,
      "developer_total_equivalent_usd": 0.03899,
      "direct_total_equivalent_usd": 0.03899,
      "actual_credit_debits": 0
    },
    {
      "task": "api-2",
      "condition": "fresh402",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 41829,
        "inputTokens": 41224,
        "cachedInputTokens": 30336,
        "cacheWriteInputTokens": 0,
        "outputTokens": 605,
        "reasoningOutputTokens": 54,
        "noncachedInputTokens": 10888,
        "reasoning_counted_in_output": true
      },
      "model_requests": 5,
      "tool_calls": 4,
      "retries": 0,
      "wall_ms": 20299,
      "inference_api_equivalent_usd": 0.030859599999999997,
      "direct_fee_equivalent_usd": 0.02,
      "prepaid_credits_equivalent": 5,
      "starter_total_equivalent_usd": 0.0433596,
      "developer_total_equivalent_usd": 0.038359599999999994,
      "direct_total_equivalent_usd": 0.0508596,
      "actual_credit_debits": 0
    },
    {
      "task": "api-3",
      "condition": "free",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 42819,
        "inputTokens": 42253,
        "cachedInputTokens": 30976,
        "cacheWriteInputTokens": 0,
        "outputTokens": 566,
        "reasoningOutputTokens": 66,
        "noncachedInputTokens": 11277,
        "reasoning_counted_in_output": true
      },
      "model_requests": 5,
      "tool_calls": 4,
      "retries": 0,
      "wall_ms": 24328,
      "inference_api_equivalent_usd": 0.0313116,
      "direct_fee_equivalent_usd": 0,
      "prepaid_credits_equivalent": 0,
      "starter_total_equivalent_usd": 0.0313116,
      "developer_total_equivalent_usd": 0.0313116,
      "direct_total_equivalent_usd": 0.0313116,
      "actual_credit_debits": 0
    },
    {
      "task": "api-3",
      "condition": "fresh402",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 50565,
        "inputTokens": 49984,
        "cachedInputTokens": 27904,
        "cacheWriteInputTokens": 0,
        "outputTokens": 581,
        "reasoningOutputTokens": 20,
        "noncachedInputTokens": 22080,
        "reasoning_counted_in_output": true
      },
      "model_requests": 5,
      "tool_calls": 4,
      "retries": 0,
      "wall_ms": 22675,
      "inference_api_equivalent_usd": 0.0527604,
      "direct_fee_equivalent_usd": 0.02,
      "prepaid_credits_equivalent": 5,
      "starter_total_equivalent_usd": 0.0652604,
      "developer_total_equivalent_usd": 0.0602604,
      "direct_total_equivalent_usd": 0.0727604,
      "actual_credit_debits": 0
    },
    {
      "task": "api-4",
      "condition": "free",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 50482,
        "inputTokens": 49942,
        "cachedInputTokens": 34560,
        "cacheWriteInputTokens": 0,
        "outputTokens": 540,
        "reasoningOutputTokens": 23,
        "noncachedInputTokens": 15382,
        "reasoning_counted_in_output": true
      },
      "model_requests": 5,
      "tool_calls": 4,
      "retries": 0,
      "wall_ms": 22463,
      "inference_api_equivalent_usd": 0.03962,
      "direct_fee_equivalent_usd": 0,
      "prepaid_credits_equivalent": 0,
      "starter_total_equivalent_usd": 0.03962,
      "developer_total_equivalent_usd": 0.03962,
      "direct_total_equivalent_usd": 0.03962,
      "actual_credit_debits": 0
    },
    {
      "task": "api-4",
      "condition": "fresh402",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 42178,
        "inputTokens": 41624,
        "cachedInputTokens": 30464,
        "cacheWriteInputTokens": 0,
        "outputTokens": 554,
        "reasoningOutputTokens": 38,
        "noncachedInputTokens": 11160,
        "reasoning_counted_in_output": true
      },
      "model_requests": 5,
      "tool_calls": 4,
      "retries": 0,
      "wall_ms": 23343,
      "inference_api_equivalent_usd": 0.0309064,
      "direct_fee_equivalent_usd": 0.02,
      "prepaid_credits_equivalent": 5,
      "starter_total_equivalent_usd": 0.0434064,
      "developer_total_equivalent_usd": 0.0384064,
      "direct_total_equivalent_usd": 0.050906400000000004,
      "actual_credit_debits": 0
    },
    {
      "task": "api-5",
      "condition": "free",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 42899,
        "inputTokens": 42338,
        "cachedInputTokens": 23296,
        "cacheWriteInputTokens": 0,
        "outputTokens": 561,
        "reasoningOutputTokens": 51,
        "noncachedInputTokens": 19042,
        "reasoning_counted_in_output": true
      },
      "model_requests": 5,
      "tool_calls": 4,
      "retries": 0,
      "wall_ms": 21502,
      "inference_api_equivalent_usd": 0.0460236,
      "direct_fee_equivalent_usd": 0,
      "prepaid_credits_equivalent": 0,
      "starter_total_equivalent_usd": 0.0460236,
      "developer_total_equivalent_usd": 0.0460236,
      "direct_total_equivalent_usd": 0.0460236,
      "actual_credit_debits": 0
    },
    {
      "task": "api-5",
      "condition": "fresh402",
      "model": "gpt-6.1-sol",
      "passed": true,
      "usage": {
        "totalTokens": 42537,
        "inputTokens": 42021,
        "cachedInputTokens": 20096,
        "cacheWriteInputTokens": 0,
        "outputTokens": 516,
        "reasoningOutputTokens": 20,
        "noncachedInputTokens": 21925,
        "reasoning_counted_in_output": true
      },
      "model_requests": 4,
      "tool_calls": 3,
      "retries": 0,
      "wall_ms": 21388,
      "inference_api_equivalent_usd": 0.0510196,
      "direct_fee_equivalent_usd": 0.02,
      "prepaid_credits_equivalent": 5,
      "starter_total_equivalent_usd": 0.0635196,
      "developer_total_equivalent_usd": 0.0585196,
      "direct_total_equivalent_usd": 0.0710196,
      "actual_credit_debits": 0
    }
  ]
}