{
 "note": "Imported provider-explorer receipts from 2026-10-03. Prompts and outputs are not copied. A cohort is a set of runs that shared a protocol. Only classification \"evaluated\" counts toward comparisons.",
 "cases": [
  {
   "key": "case-1",
   "title": "Fixed exact reply",
   "kind": "latency",
   "description": "Exact short reply used by the historical latency cohorts."
  },
  {
   "key": "case-2",
   "title": "Fixed 243-token answer",
   "kind": "latency",
   "description": "Fixed 48-line answer with 243 output tokens."
  },
  {
   "key": "case-3",
   "title": "Synthetic coding task (API prompt)",
   "kind": "coding",
   "description": "Synthetic coding task used by direct API rows and API controls."
  },
  {
   "key": "case-4",
   "title": "Synthetic coding task (CLI matched prompt)",
   "kind": "coding",
   "description": "Synthetic coding task used by matched and profile CLI rows."
  },
  {
   "key": "case-5",
   "title": "Synthetic coding task (CLI parity prompt)",
   "kind": "coding",
   "description": "Synthetic coding task used by parity and captured-wire CLI rows."
  },
  {
   "key": "case-6",
   "title": "PlanJobs scheduler repair",
   "kind": "coding",
   "description": "Dependency-aware planJobs scheduler repair task."
  },
  {
   "key": "case-7",
   "title": "Context inspection diagnostic",
   "kind": "diagnostic",
   "description": "Mock request context inspection without provider inference."
  },
  {
   "key": "case-8",
   "title": "Scratch file edit probe",
   "kind": "coding",
   "description": "Actual scratch-file editing smoke probe."
  },
  {
   "key": "case-9",
   "title": "mergeRanges coding repair",
   "kind": "coding",
   "description": "mergeRanges repair task used by the followup and Flex coding probes."
  },
  {
   "key": "case-10",
   "title": "CLI startup diagnostic",
   "kind": "diagnostic",
   "description": "CLI startup, event semantics, and eligibility probe."
  },
  {
   "key": "case-11",
   "title": "Model catalog diagnostic",
   "kind": "diagnostic",
   "description": "Model catalog and account-route capability probe."
  }
 ],
 "runs": [
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2933,
   "totalMs": 3302,
   "inputTokens": 8295,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 2688,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2579,
   "totalMs": 7760,
   "inputTokens": 4945,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.014791,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 886,
   "totalMs": 3631,
   "inputTokens": 363,
   "outputTokens": 232,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.003046,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 12896,
   "totalMs": 13492,
   "inputTokens": 26510,
   "outputTokens": 259,
   "reasoningTokens": 0,
   "cacheReadTokens": 12928,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-5",
   "caseTitle": "Synthetic coding task (CLI parity prompt)",
   "caseKind": "coding",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3841,
   "totalMs": 4046,
   "inputTokens": 8492,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0233285,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-5",
   "caseTitle": "Synthetic coding task (CLI parity prompt)",
   "caseKind": "coding",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4253,
   "totalMs": 4463,
   "inputTokens": 8496,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0233385,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 6416,
   "totalMs": 10940,
   "inputTokens": 382,
   "outputTokens": 450,
   "reasoningTokens": 191,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.005264,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 11850,
   "totalMs": 12428,
   "inputTokens": 18288,
   "outputTokens": 239,
   "reasoningTokens": 0,
   "cacheReadTokens": 8832,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-initial",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1289,
   "totalMs": 4061,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 848,
   "totalMs": 959,
   "inputTokens": 166,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000402,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4993,
   "totalMs": 9564,
   "inputTokens": 382,
   "outputTokens": 429,
   "reasoningTokens": 155,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.005054,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3339,
   "totalMs": 12056,
   "inputTokens": 6620,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2581,
   "totalMs": 2973,
   "inputTokens": 10996,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0275585,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 554,
   "totalMs": 713,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0000052,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "api-initial",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1781,
   "totalMs": 1895,
   "inputTokens": 17,
   "outputTokens": 17,
   "reasoningTokens": 8,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000204,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3750,
   "totalMs": 4183,
   "inputTokens": 19551,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 12288,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "bare-api",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 990,
   "totalMs": 3921,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "native-pilot",
   "classification": "excluded",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 5169,
   "inputTokens": 4949,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.014801,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2888,
   "totalMs": 10689,
   "inputTokens": 11258,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "claude-scheduler",
   "classification": "evaluated",
   "status": "completed",
   "route": "claude-cli",
   "model": "claude-sonnet-5-5",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": "2.1.284 (Claude Code)",
   "firstUsefulMs": 7554,
   "totalMs": 14998,
   "inputTokens": 2,
   "outputTokens": 2227,
   "reasoningTokens": null,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": 0.030119,
   "costStatus": "list-price",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2328,
   "totalMs": 2758,
   "inputTokens": 6360,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0159685,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 12520,
   "totalMs": 13023,
   "inputTokens": 39529,
   "outputTokens": 264,
   "reasoningTokens": 0,
   "cacheReadTokens": 31744,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 744,
   "totalMs": 3529,
   "inputTokens": 363,
   "outputTokens": 238,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.003106,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2787,
   "totalMs": 3192,
   "inputTokens": 18859,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 11008,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1637,
   "totalMs": 4962,
   "inputTokens": 6624,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0189885,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2900,
   "totalMs": 3312,
   "inputTokens": 8295,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 2688,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "native",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 12513,
   "inputTokens": 4945,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler-context-mismatch",
   "classification": "excluded",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 14893,
   "totalMs": 23864,
   "inputTokens": 5461,
   "outputTokens": 1317,
   "reasoningTokens": 292,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": 0.0134105,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3435,
   "totalMs": 3860,
   "inputTokens": 19551,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 12288,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-8",
   "caseTitle": "Scratch file edit probe",
   "caseKind": "coding",
   "cohort": "native-pilot",
   "classification": "excluded",
   "status": "failed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 39189,
   "inputTokens": 16170,
   "outputTokens": 340,
   "reasoningTokens": 41,
   "cacheReadTokens": 4992,
   "passed": false,
   "checks": 1,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 5695,
   "totalMs": 6076,
   "inputTokens": 22387,
   "outputTokens": 261,
   "reasoningTokens": 0,
   "cacheReadTokens": 11100,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0319345,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 17271,
   "totalMs": 17847,
   "inputTokens": 39537,
   "outputTokens": 405,
   "reasoningTokens": 107,
   "cacheReadTokens": 33920,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 13647,
   "totalMs": 59897,
   "inputTokens": 9563,
   "outputTokens": 1181,
   "reasoningTokens": 135,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1742,
   "totalMs": 1865,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000104,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "flex-initial",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 3299,
   "totalMs": 5569,
   "inputTokens": 206,
   "outputTokens": 289,
   "reasoningTokens": 51,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.001651,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4988,
   "totalMs": 5383,
   "inputTokens": 8937,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 3328,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2316,
   "totalMs": 2755,
   "inputTokens": 6358,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0159635,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 944,
   "totalMs": 3679,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4296,
   "totalMs": 4691,
   "inputTokens": 19553,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 12288,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 13599,
   "totalMs": 14149,
   "inputTokens": 39532,
   "outputTokens": 266,
   "reasoningTokens": 0,
   "cacheReadTokens": 31744,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2031,
   "totalMs": 2464,
   "inputTokens": 6360,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0159685,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 1184,
   "totalMs": 3222,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001494,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4236,
   "totalMs": 4580,
   "inputTokens": 8295,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 2688,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2939,
   "totalMs": 3346,
   "inputTokens": 18859,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 11008,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3257,
   "totalMs": 3722,
   "inputTokens": 8299,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 2688,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 624,
   "totalMs": 4350,
   "inputTokens": 382,
   "outputTokens": 226,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0001512,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1098,
   "totalMs": 3802,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "native-pilot",
   "classification": "excluded",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 12547,
   "inputTokens": 4945,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2491,
   "totalMs": 2907,
   "inputTokens": 18859,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 7936,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1201,
   "totalMs": 3766,
   "inputTokens": 8496,
   "outputTokens": 225,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0234885,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1670,
   "totalMs": 1922,
   "inputTokens": 8297,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020811,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3241,
   "totalMs": 3635,
   "inputTokens": 13042,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "bare-api",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1619,
   "totalMs": 4452,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-5",
   "caseTitle": "Synthetic coding task (CLI parity prompt)",
   "caseKind": "coding",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4860,
   "totalMs": 5085,
   "inputTokens": 8498,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0233435,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3299,
   "totalMs": 14947,
   "inputTokens": 6549,
   "outputTokens": 225,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 8274,
   "totalMs": 8992,
   "inputTokens": 38110,
   "outputTokens": 285,
   "reasoningTokens": 0,
   "cacheReadTokens": 29184,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4103,
   "totalMs": 4532,
   "inputTokens": 19551,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 12288,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "setup-failures",
   "classification": "diagnostic",
   "status": "failed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": false,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3753,
   "totalMs": 4127,
   "inputTokens": 19553,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 12288,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 13199,
   "totalMs": 13810,
   "inputTokens": 18271,
   "outputTokens": 246,
   "reasoningTokens": 0,
   "cacheReadTokens": 8832,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 991,
   "totalMs": 3051,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001494,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 5502,
   "totalMs": 12181,
   "inputTokens": 4874,
   "outputTokens": 241,
   "reasoningTokens": 29,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler-excluded",
   "classification": "excluded",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 12155,
   "totalMs": 20235,
   "inputTokens": 5461,
   "outputTokens": 1330,
   "reasoningTokens": 300,
   "cacheReadTokens": 5458,
   "passed": true,
   "checks": 296,
   "costUsd": 0.0069259,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3194,
   "totalMs": 3942,
   "inputTokens": 10993,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 5312,
   "totalMs": 9444,
   "inputTokens": 382,
   "outputTokens": 429,
   "reasoningTokens": 170,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.005054,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2045,
   "totalMs": 5880,
   "inputTokens": 4949,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.014801,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4383,
   "totalMs": 12661,
   "inputTokens": 4947,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2074,
   "totalMs": 2163,
   "inputTokens": 166,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000402,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-5",
   "caseTitle": "Synthetic coding task (CLI parity prompt)",
   "caseKind": "coding",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 9414,
   "totalMs": 9885,
   "inputTokens": 8492,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 7891,
   "totalMs": 8327,
   "inputTokens": 22388,
   "outputTokens": 239,
   "reasoningTokens": 0,
   "cacheReadTokens": 11100,
   "passed": true,
   "checks": 120,
   "costUsd": 0.031717,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-5",
   "caseTitle": "Synthetic coding task (CLI parity prompt)",
   "caseKind": "coding",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 9622,
   "totalMs": 10054,
   "inputTokens": 8496,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 2688,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4977,
   "totalMs": 12560,
   "inputTokens": 6549,
   "outputTokens": 267,
   "reasoningTokens": 35,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1406,
   "totalMs": 4224,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1503,
   "totalMs": 4006,
   "inputTokens": 8496,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0233385,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 17127,
   "totalMs": 17681,
   "inputTokens": 39496,
   "outputTokens": 394,
   "reasoningTokens": 90,
   "cacheReadTokens": 31744,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1259,
   "totalMs": 1348,
   "inputTokens": 17,
   "outputTokens": 17,
   "reasoningTokens": 8,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000204,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-10",
   "caseTitle": "CLI startup diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3373,
   "totalMs": 3838,
   "inputTokens": 8935,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-8",
   "caseTitle": "Scratch file edit probe",
   "caseKind": "coding",
   "cohort": "edit-check",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 8339,
   "inputTokens": 35281,
   "outputTokens": 323,
   "reasoningTokens": 0,
   "cacheReadTokens": 23225,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "api"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 9272,
   "totalMs": 9905,
   "inputTokens": 18236,
   "outputTokens": 261,
   "reasoningTokens": 0,
   "cacheReadTokens": 7936,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1371,
   "totalMs": 1502,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0000052,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 873,
   "totalMs": 1021,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000104,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-5",
   "caseTitle": "Synthetic coding task (CLI parity prompt)",
   "caseKind": "coding",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 12150,
   "totalMs": 12516,
   "inputTokens": 8492,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 2688,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3889,
   "totalMs": 4349,
   "inputTokens": 19551,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 14464,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 5555,
   "totalMs": 12064,
   "inputTokens": 6551,
   "outputTokens": 251,
   "reasoningTokens": 39,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1571,
   "totalMs": 4800,
   "inputTokens": 11266,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0305935,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler-context-mismatch",
   "classification": "excluded",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 15560,
   "totalMs": 61164,
   "inputTokens": 9563,
   "outputTokens": 1163,
   "reasoningTokens": 156,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3010,
   "totalMs": 6236,
   "inputTokens": 6547,
   "outputTokens": 263,
   "reasoningTokens": 36,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.018996,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2766,
   "totalMs": 3672,
   "inputTokens": 10998,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 974,
   "totalMs": 1106,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000104,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3473,
   "totalMs": 3847,
   "inputTokens": 8935,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3472,
   "totalMs": 10075,
   "inputTokens": 4874,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "claude-scheduler",
   "classification": "evaluated",
   "status": "completed",
   "route": "claude-cli",
   "model": "claude-sonnet-5-5",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": "2.1.284 (Claude Code)",
   "firstUsefulMs": 6768,
   "totalMs": 13885,
   "inputTokens": 2,
   "outputTokens": 2144,
   "reasoningTokens": null,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": 0.029289000000000003,
   "costStatus": "list-price",
   "billing": "subscription"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler-excluded",
   "classification": "excluded",
   "status": "failed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 5136,
   "totalMs": 5913,
   "inputTokens": 10991,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 823,
   "totalMs": 965,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0000052,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3426,
   "totalMs": 4276,
   "inputTokens": 10993,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 7424,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 7557,
   "totalMs": 8165,
   "inputTokens": 18244,
   "outputTokens": 241,
   "reasoningTokens": 0,
   "cacheReadTokens": 7936,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1479,
   "totalMs": 1609,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000104,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-8",
   "caseTitle": "Scratch file edit probe",
   "caseKind": "coding",
   "cohort": "native",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 11147,
   "inputTokens": 18236,
   "outputTokens": 406,
   "reasoningTokens": 0,
   "cacheReadTokens": 11523,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0219903,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-11",
   "caseTitle": "Model catalog diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1089,
   "totalMs": 1193,
   "inputTokens": 166,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000402,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2461,
   "totalMs": 2875,
   "inputTokens": 18861,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 11008,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 13270,
   "totalMs": 13816,
   "inputTokens": 26507,
   "outputTokens": 286,
   "reasoningTokens": 0,
   "cacheReadTokens": 12928,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3780,
   "totalMs": 12450,
   "inputTokens": 11262,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-5",
   "caseTitle": "Synthetic coding task (CLI parity prompt)",
   "caseKind": "coding",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3896,
   "totalMs": 4086,
   "inputTokens": 8494,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0233335,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1054,
   "totalMs": 5996,
   "inputTokens": 382,
   "outputTokens": 232,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.003084,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1467,
   "totalMs": 4826,
   "inputTokens": 6622,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0189835,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 506,
   "totalMs": 645,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0000052,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-initial",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 2464,
   "totalMs": 4804,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001494,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 1083,
   "totalMs": 3242,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001494,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 13637,
   "totalMs": 14421,
   "inputTokens": 22391,
   "outputTokens": 246,
   "reasoningTokens": 0,
   "cacheReadTokens": 10880,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2921,
   "totalMs": 10923,
   "inputTokens": 4945,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1746,
   "totalMs": 2161,
   "inputTokens": 10993,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.027551,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 1187,
   "totalMs": 3330,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001494,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 23035,
   "totalMs": 69512,
   "inputTokens": 9563,
   "outputTokens": 1295,
   "reasoningTokens": 254,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 8681,
   "totalMs": 9231,
   "inputTokens": 38115,
   "outputTokens": 292,
   "reasoningTokens": 0,
   "cacheReadTokens": 29184,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1936,
   "totalMs": 2343,
   "inputTokens": 10998,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0275635,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 2096,
   "totalMs": 4349,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001494,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1126,
   "totalMs": 3932,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler-context-mismatch",
   "classification": "excluded",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 13426,
   "totalMs": 22729,
   "inputTokens": 5461,
   "outputTokens": 1308,
   "reasoningTokens": 272,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": 0.0133655,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 814,
   "totalMs": 4010,
   "inputTokens": 382,
   "outputTokens": 226,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0001512,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 967,
   "totalMs": 5439,
   "inputTokens": 382,
   "outputTokens": 257,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.003334,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "wire-failure",
   "classification": "excluded",
   "status": "failed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": null,
   "totalMs": 51,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": false,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3808,
   "totalMs": 6272,
   "inputTokens": 8492,
   "outputTokens": 258,
   "reasoningTokens": 46,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0238085,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "native-pilot",
   "classification": "excluded",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 12062,
   "inputTokens": 4945,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2573,
   "totalMs": 2843,
   "inputTokens": 8295,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020806,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-5",
   "caseTitle": "Synthetic coding task (CLI parity prompt)",
   "caseKind": "coding",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3538,
   "totalMs": 3731,
   "inputTokens": 8496,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0233385,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3364,
   "totalMs": 11202,
   "inputTokens": 11260,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "excluded",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 2037,
   "totalMs": 4230,
   "inputTokens": 279,
   "outputTokens": 257,
   "reasoningTokens": 12,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001564,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "native",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 5587,
   "inputTokens": 4945,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.014791,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 21860,
   "totalMs": 22418,
   "inputTokens": 39527,
   "outputTokens": 462,
   "reasoningTokens": 164,
   "cacheReadTokens": 31744,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "native",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 5144,
   "inputTokens": 4947,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.014796,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2187,
   "totalMs": 2433,
   "inputTokens": 8297,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020811,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3663,
   "totalMs": 12180,
   "inputTokens": 4945,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3406,
   "totalMs": 3840,
   "inputTokens": 23143,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 12288,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "failed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1792,
   "totalMs": 2079,
   "inputTokens": 8295,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020806,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 10081,
   "totalMs": 10565,
   "inputTokens": 18244,
   "outputTokens": 281,
   "reasoningTokens": 0,
   "cacheReadTokens": 7936,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2533,
   "totalMs": 2760,
   "inputTokens": 8297,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020811,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1398,
   "totalMs": 6197,
   "inputTokens": 382,
   "outputTokens": 238,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.003144,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 12864,
   "totalMs": 13685,
   "inputTokens": 22391,
   "outputTokens": 256,
   "reasoningTokens": 0,
   "cacheReadTokens": 10880,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "native",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 12131,
   "inputTokens": 4945,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 2304,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler-context-mismatch",
   "classification": "excluded",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 11284,
   "totalMs": 20850,
   "inputTokens": 5461,
   "outputTokens": 1254,
   "reasoningTokens": 227,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": 0.0130955,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1414,
   "totalMs": 4223,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 5719,
   "totalMs": 13517,
   "inputTokens": 6620,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1341,
   "totalMs": 1420,
   "inputTokens": 17,
   "outputTokens": 17,
   "reasoningTokens": 8,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000204,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1318,
   "totalMs": 1520,
   "inputTokens": 17,
   "outputTokens": 17,
   "reasoningTokens": 8,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000204,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4006,
   "totalMs": 4375,
   "inputTokens": 8297,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3063,
   "totalMs": 6021,
   "inputTokens": 6549,
   "outputTokens": 253,
   "reasoningTokens": 41,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.018901,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4504,
   "totalMs": 6589,
   "inputTokens": 6360,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "bare-api",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1373,
   "totalMs": 4112,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-initial",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 1383,
   "totalMs": 3394,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001494,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3529,
   "totalMs": 3970,
   "inputTokens": 13042,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-8",
   "caseTitle": "Scratch file edit probe",
   "caseKind": "coding",
   "cohort": "installed-live",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 19347,
   "inputTokens": 18131,
   "outputTokens": 350,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "setup-failures",
   "classification": "diagnostic",
   "status": "failed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": false,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1584,
   "totalMs": 4856,
   "inputTokens": 11262,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0305835,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 15857,
   "totalMs": 16412,
   "inputTokens": 26513,
   "outputTokens": 298,
   "reasoningTokens": 0,
   "cacheReadTokens": 12928,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler-context-mismatch",
   "classification": "excluded",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 13647,
   "totalMs": 59897,
   "inputTokens": 9563,
   "outputTokens": 1181,
   "reasoningTokens": 135,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1581,
   "totalMs": 1700,
   "inputTokens": 8295,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020806,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "claude-scheduler",
   "classification": "evaluated",
   "status": "completed",
   "route": "claude-cli",
   "model": "claude-sonnet-5-5",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": "2.1.284 (Claude Code)",
   "firstUsefulMs": 7633,
   "totalMs": 15894,
   "inputTokens": 2,
   "outputTokens": 2303,
   "reasoningTokens": null,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": 0.0308815,
   "costStatus": "list-price",
   "billing": "subscription"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 6678,
   "totalMs": 16280,
   "inputTokens": 9563,
   "outputTokens": 1313,
   "reasoningTokens": 267,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": 0.018518,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 13832,
   "totalMs": 14411,
   "inputTokens": 39527,
   "outputTokens": 266,
   "reasoningTokens": 0,
   "cacheReadTokens": 31744,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2115,
   "totalMs": 2233,
   "inputTokens": 17,
   "outputTokens": 17,
   "reasoningTokens": 8,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000204,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1980,
   "totalMs": 2349,
   "inputTokens": 10993,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.027551,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-initial",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1143,
   "totalMs": 3839,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-initial",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 1633,
   "totalMs": 3690,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001494,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "native-pilot",
   "classification": "excluded",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 5480,
   "inputTokens": 4945,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.014791,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 818,
   "totalMs": 3774,
   "inputTokens": 363,
   "outputTokens": 242,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.003146,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 848,
   "totalMs": 1098,
   "inputTokens": 166,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000402,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 838,
   "totalMs": 1057,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0000052,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-10",
   "caseTitle": "CLI startup diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1876,
   "totalMs": 4827,
   "inputTokens": 4872,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0142785,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler-context-mismatch",
   "classification": "excluded",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 23035,
   "totalMs": 69512,
   "inputTokens": 9563,
   "outputTokens": 1295,
   "reasoningTokens": 254,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3826,
   "totalMs": 4258,
   "inputTokens": 23141,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 12288,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1882,
   "totalMs": 5730,
   "inputTokens": 4943,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.014786,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1154,
   "totalMs": 1341,
   "inputTokens": 8297,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020811,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-8",
   "caseTitle": "Scratch file edit probe",
   "caseKind": "coding",
   "cohort": "installed-live",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 8633,
   "inputTokens": 18156,
   "outputTokens": 345,
   "reasoningTokens": 0,
   "cacheReadTokens": 11507,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0212187,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-8",
   "caseTitle": "Scratch file edit probe",
   "caseKind": "coding",
   "cohort": "native",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": 20159,
   "inputTokens": 18185,
   "outputTokens": 408,
   "reasoningTokens": 0,
   "cacheReadTokens": 11136,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2008,
   "totalMs": 5171,
   "inputTokens": 11260,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0305785,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3246,
   "totalMs": 6175,
   "inputTokens": 6551,
   "outputTokens": 258,
   "reasoningTokens": 46,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.018956,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1442,
   "totalMs": 1647,
   "inputTokens": 8299,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020816,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-initial",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 2399,
   "totalMs": 5189,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 936,
   "totalMs": 1089,
   "inputTokens": 166,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000402,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "unsupported-control",
   "classification": "excluded",
   "status": "failed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.157.0",
   "firstUsefulMs": null,
   "totalMs": 2672,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": false,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-5",
   "caseTitle": "Synthetic coding task (CLI parity prompt)",
   "caseKind": "coding",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4177,
   "totalMs": 4420,
   "inputTokens": 8496,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0233385,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2962,
   "totalMs": 6078,
   "inputTokens": 4872,
   "outputTokens": 241,
   "reasoningTokens": 29,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0145885,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-11",
   "caseTitle": "Model catalog diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-minimal",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3281,
   "totalMs": 6095,
   "inputTokens": 4872,
   "outputTokens": 255,
   "reasoningTokens": 43,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0147285,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-7",
   "caseTitle": "Context inspection diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "failed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 6835,
   "totalMs": 7277,
   "inputTokens": 22395,
   "outputTokens": 277,
   "reasoningTokens": 29,
   "cacheReadTokens": 11104,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0321049,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3021,
   "totalMs": 3733,
   "inputTokens": 6358,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 3328,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 1083,
   "totalMs": 3765,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "flex-initial",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 3159,
   "totalMs": 5979,
   "inputTokens": 206,
   "outputTokens": 300,
   "reasoningTokens": 58,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.003412,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "excluded",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 2037,
   "totalMs": 4740,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.002988,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2656,
   "totalMs": 2890,
   "inputTokens": 8295,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020806,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 7460,
   "totalMs": 17317,
   "inputTokens": 9563,
   "outputTokens": 1263,
   "reasoningTokens": 214,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": 0.018268,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "flex-warm-balanced",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "flex",
   "cliVersion": null,
   "firstUsefulMs": 1273,
   "totalMs": 3751,
   "inputTokens": 279,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.001494,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 4764,
   "totalMs": 5135,
   "inputTokens": 13042,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2056,
   "totalMs": 2445,
   "inputTokens": 10991,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.027546,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3024,
   "totalMs": 4426,
   "inputTokens": 11004,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-10",
   "caseTitle": "CLI startup diagnostic",
   "caseKind": "diagnostic",
   "cohort": "diagnostics",
   "classification": "diagnostic",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": null,
   "totalMs": null,
   "inputTokens": null,
   "outputTokens": null,
   "reasoningTokens": null,
   "cacheReadTokens": null,
   "passed": null,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "unknown"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2764,
   "totalMs": 3173,
   "inputTokens": 11000,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0275685,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-3",
   "caseTitle": "Synthetic coding task (API prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 674,
   "totalMs": 3825,
   "inputTokens": 382,
   "outputTokens": 226,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": 0.0001512,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 835,
   "totalMs": 957,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000104,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "wire-matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1960,
   "totalMs": 2089,
   "inputTokens": 8299,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020816,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 15560,
   "totalMs": 61164,
   "inputTokens": 9563,
   "outputTokens": 1163,
   "reasoningTokens": 156,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3786,
   "totalMs": 4191,
   "inputTokens": 19555,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 12288,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "installed",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 11989,
   "totalMs": 15333,
   "inputTokens": 22390,
   "outputTokens": 246,
   "reasoningTokens": 0,
   "cacheReadTokens": 10880,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 1847,
   "totalMs": 5168,
   "inputTokens": 6622,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.0189835,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-6",
   "caseTitle": "PlanJobs scheduler repair",
   "caseKind": "coding",
   "cohort": "scheduler",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "medium",
   "serviceTier": "standard",
   "cliVersion": null,
   "firstUsefulMs": 9049,
   "totalMs": 18612,
   "inputTokens": 9563,
   "outputTokens": 1388,
   "reasoningTokens": 349,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 296,
   "costUsd": 0.018893,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "openai-api",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 841,
   "totalMs": 1008,
   "inputTokens": 17,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.000104,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "parity",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 2091,
   "totalMs": 2296,
   "inputTokens": 8299,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": 0.020816,
   "costStatus": "estimate",
   "billing": "api"
  },
  {
   "case": "case-9",
   "caseTitle": "mergeRanges coding repair",
   "caseKind": "coding",
   "cohort": "followup-minimal",
   "classification": "excluded",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 9160,
   "totalMs": 21235,
   "inputTokens": 4874,
   "outputTokens": 210,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3331,
   "totalMs": 4033,
   "inputTokens": 11000,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 6934,
   "totalMs": 7627,
   "inputTokens": 6360,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 11012,
   "totalMs": 11679,
   "inputTokens": 38101,
   "outputTokens": 280,
   "reasoningTokens": 0,
   "cacheReadTokens": 36352,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-4",
   "caseTitle": "Synthetic coding task (CLI matched prompt)",
   "caseKind": "coding",
   "cohort": "tuning",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 12793,
   "totalMs": 13450,
   "inputTokens": 18279,
   "outputTokens": 261,
   "reasoningTokens": 0,
   "cacheReadTokens": 8832,
   "passed": true,
   "checks": 120,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-2",
   "caseTitle": "Fixed 243-token answer",
   "caseKind": "latency",
   "cohort": "followup-stream",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "low",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3744,
   "totalMs": 11576,
   "inputTokens": 6622,
   "outputTokens": 243,
   "reasoningTokens": 0,
   "cacheReadTokens": 0,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6.1-sol",
   "effort": "high",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3366,
   "totalMs": 3810,
   "inputTokens": 19553,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 12288,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  },
  {
   "case": "case-1",
   "caseTitle": "Fixed exact reply",
   "caseKind": "latency",
   "cohort": "matched",
   "classification": "evaluated",
   "status": "completed",
   "route": "codex-cli",
   "model": "gpt-6-luna",
   "effort": "none",
   "serviceTier": "standard",
   "cliVersion": "0.160.0",
   "firstUsefulMs": 3418,
   "totalMs": 3831,
   "inputTokens": 18863,
   "outputTokens": 7,
   "reasoningTokens": 0,
   "cacheReadTokens": 11008,
   "passed": true,
   "checks": null,
   "costUsd": null,
   "costStatus": "unknown",
   "billing": "subscription"
  }
 ]
}