{
  "experiment_date": "2026-09-17",
  "exchange_rate_assumption_krw_per_usd": 1400,
  "comparison": {
    "live": {
      "n": 4,
      "state_correct": 4,
      "state_resolved": 3,
      "state_handoff": 1,
      "model_finished": 4,
      "usd_low": 0.12582975,
      "usd_high": 0.151278,
      "mean_usd_low": 0.0314574375,
      "mean_usd_high": 0.0378195,
      "first_audio_s_median": 4.995379354048055,
      "action_complete_s_median": 23.736457083576358,
      "correct_action_complete_s_median": 23.736457083576358
    },
    "extended": {
      "n": 4,
      "state_correct": 4,
      "state_resolved": 3,
      "state_handoff": 1,
      "model_finished": 2,
      "usd_low": 0.7087282500000001,
      "usd_high": 0.72176025,
      "mean_usd_low": 0.17718206250000001,
      "mean_usd_high": 0.1804400625,
      "first_audio_s_median": 1.357734728953801,
      "action_complete_s_median": 97.23350089599518,
      "correct_action_complete_s_median": 97.23350089599518
    }
  },
  "excluded_cases": {
    "ambiguous_budget": "Live registered budget_scope but audibly asked to increase budget; synthetic caller supplied scope information. Exclude both sides from paired outcome/cost/time comparison.",
    "one_seat": "Extended caller answered the registered question before its spoken wording appeared; exclude both sides from cost/time comparison."
  },
  "rows": [
    {
      "run": "main-ambiguous_budget-extended-r1",
      "case": "ambiguous_budget",
      "model": "extended",
      "repeat": 1,
      "stop": "session_timeout",
      "model_finished": false,
      "model_complete_s": null,
      "correct": false,
      "completed_correct": false,
      "resolved": false,
      "handoff": false,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 1.6139561249874532,
      "valid_proposal_s": null,
      "action_complete_s": null,
      "duration_s": 150.89,
      "usd_low": 0.18376875,
      "usd_high": 0.18618075,
      "raw_usd_low": 0.18376875,
      "raw_usd_high": 0.18618075,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-ambiguous_budget-live-r1",
      "case": "ambiguous_budget",
      "model": "live",
      "repeat": 1,
      "stop": "completed",
      "model_finished": true,
      "model_complete_s": 44.55013712494355,
      "correct": true,
      "completed_correct": true,
      "resolved": true,
      "handoff": false,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 4.796792750013992,
      "valid_proposal_s": 18.281137124943548,
      "action_complete_s": 35.657137124943546,
      "duration_s": 62.905,
      "usd_low": 0.04290299999999999,
      "usd_high": 0.05147324999999999,
      "raw_usd_low": 0.04290299999999999,
      "raw_usd_high": 0.05147324999999999,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-later_meeting-extended-r1",
      "case": "later_meeting",
      "model": "extended",
      "repeat": 1,
      "stop": "session_timeout",
      "model_finished": false,
      "model_complete_s": null,
      "correct": true,
      "completed_correct": false,
      "resolved": true,
      "handoff": false,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 1.2715254579670727,
      "valid_proposal_s": 16.130835541976616,
      "action_complete_s": 112.63783554197661,
      "duration_s": 150.906,
      "usd_low": 0.19058325,
      "usd_high": 0.193839,
      "raw_usd_low": 0.19058325,
      "raw_usd_high": 0.193839,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-later_meeting-live-r1",
      "case": "later_meeting",
      "model": "live",
      "repeat": 1,
      "stop": "completed",
      "model_finished": true,
      "model_complete_s": 31.700379500061274,
      "correct": true,
      "completed_correct": true,
      "resolved": true,
      "handoff": false,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 4.606537415995263,
      "valid_proposal_s": 3.3093795000612722,
      "action_complete_s": 23.542379500061273,
      "duration_s": 50.686,
      "usd_low": 0.03057675,
      "usd_high": 0.036492,
      "raw_usd_low": 0.03057675,
      "raw_usd_high": 0.036492,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-one_seat-extended-r1",
      "case": "one_seat",
      "model": "extended",
      "repeat": 1,
      "stop": "completed",
      "model_finished": true,
      "model_complete_s": 69.93352045795322,
      "correct": true,
      "completed_correct": true,
      "resolved": false,
      "handoff": true,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 1.2820854999590665,
      "valid_proposal_s": null,
      "action_complete_s": 38.457520457953215,
      "duration_s": 88.985,
      "usd_low": 0.13879125,
      "usd_high": 0.1410795,
      "raw_usd_low": 0.13879125,
      "raw_usd_high": 0.1410795,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-one_seat-live-r1",
      "case": "one_seat",
      "model": "live",
      "repeat": 1,
      "stop": "completed",
      "model_finished": true,
      "model_complete_s": 30.915299584049734,
      "correct": true,
      "completed_correct": true,
      "resolved": false,
      "handoff": true,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 7.671689417096786,
      "valid_proposal_s": null,
      "action_complete_s": 24.958299584049733,
      "duration_s": 49.853,
      "usd_low": 0.03572175,
      "usd_high": 0.04335075,
      "raw_usd_low": 0.03572175,
      "raw_usd_high": 0.04335075,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-operations-extended-r1",
      "case": "operations",
      "model": "extended",
      "repeat": 1,
      "stop": "session_timeout",
      "model_finished": true,
      "model_complete_s": 94.74397620908078,
      "correct": true,
      "completed_correct": true,
      "resolved": true,
      "handoff": false,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 1.267327000037767,
      "valid_proposal_s": 14.57097620908078,
      "action_complete_s": 69.77797620908078,
      "duration_s": 150.917,
      "usd_low": 0.126639,
      "usd_high": 0.1291185,
      "raw_usd_low": 0.171414,
      "raw_usd_high": 0.17539200000000002,
      "extra_after_completion_usd": 0.04627350000000002
    },
    {
      "run": "main-operations-live-r1",
      "case": "operations",
      "model": "live",
      "repeat": 1,
      "stop": "completed",
      "model_finished": true,
      "model_complete_s": 32.91653466709145,
      "correct": true,
      "completed_correct": true,
      "resolved": true,
      "handoff": false,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 5.384221292100847,
      "valid_proposal_s": 3.978534667091445,
      "action_complete_s": 23.930534667091443,
      "duration_s": 51.812,
      "usd_low": 0.03432075,
      "usd_high": 0.0416415,
      "raw_usd_low": 0.03432075,
      "raw_usd_high": 0.0416415,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-raised_budget-extended-r1",
      "case": "raised_budget",
      "model": "extended",
      "repeat": 1,
      "stop": "completed",
      "model_finished": true,
      "model_complete_s": 128.7574738749815,
      "correct": true,
      "completed_correct": true,
      "resolved": true,
      "handoff": false,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 1.4439439999405295,
      "valid_proposal_s": 12.149473874981513,
      "action_complete_s": 98.69547387498152,
      "duration_s": 147.773,
      "usd_low": 0.19007175,
      "usd_high": 0.19402725,
      "raw_usd_low": 0.19007175,
      "raw_usd_high": 0.19402725,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-raised_budget-live-r1",
      "case": "raised_budget",
      "model": "live",
      "repeat": 1,
      "stop": "completed",
      "model_finished": true,
      "model_complete_s": 32.74576020895876,
      "correct": true,
      "completed_correct": true,
      "resolved": true,
      "handoff": false,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 6.446474584052339,
      "valid_proposal_s": 4.173760208958761,
      "action_complete_s": 25.34176020895876,
      "duration_s": 51.697,
      "usd_low": 0.03311325,
      "usd_high": 0.039036,
      "raw_usd_low": 0.03311325,
      "raw_usd_high": 0.039036,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-weather-extended-r1",
      "case": "weather",
      "model": "extended",
      "repeat": 1,
      "stop": "session_timeout",
      "model_finished": false,
      "model_complete_s": null,
      "correct": true,
      "completed_correct": false,
      "resolved": false,
      "handoff": true,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 1.6312836669385433,
      "valid_proposal_s": null,
      "action_complete_s": 95.77152791700885,
      "duration_s": 151.129,
      "usd_low": 0.20143425,
      "usd_high": 0.2047755,
      "raw_usd_low": 0.20143425,
      "raw_usd_high": 0.2047755,
      "extra_after_completion_usd": 0.0
    },
    {
      "run": "main-weather-live-r1",
      "case": "weather",
      "model": "live",
      "repeat": 1,
      "stop": "completed",
      "model_finished": true,
      "model_complete_s": 26.799805584017186,
      "correct": true,
      "completed_correct": true,
      "resolved": false,
      "handoff": true,
      "invalid_proposals": 0,
      "wrong_commit_attempts": 0,
      "first_audio_s": 4.585608125082217,
      "valid_proposal_s": null,
      "action_complete_s": 21.955805584017185,
      "duration_s": 45.691,
      "usd_low": 0.027819,
      "usd_high": 0.0341085,
      "raw_usd_low": 0.027819,
      "raw_usd_high": 0.0341085,
      "extra_after_completion_usd": 0.0
    }
  ],
  "expense_groups_usd": {
    "new_pilots": {
      "sessions": 9,
      "usd_low": 0.7092854999999999,
      "usd_high": 0.7500067499999999
    },
    "new_main": {
      "sessions": 12,
      "usd_low": 1.28051775,
      "usd_high": 1.341396
    },
    "new_all": {
      "sessions": 21,
      "usd_low": 1.98980325,
      "usd_high": 2.09140275
    },
    "prior_19": {
      "sessions": 19,
      "usd_low": 0.92108475,
      "usd_high": 0.9638565
    },
    "all_experiments": {
      "sessions": 40,
      "usd_low": 2.910888,
      "usd_high": 3.0552592499999998
    }
  },
  "expense_groups_krw": {
    "new_pilots": {
      "low": 992.9996999999998,
      "high": 1050.00945
    },
    "new_main": {
      "low": 1792.72485,
      "high": 1877.9544
    },
    "new_all": {
      "low": 2785.72455,
      "high": 2927.9638499999996
    },
    "prior_19": {
      "low": 1289.51865,
      "high": 1349.3991
    },
    "all_experiments": {
      "low": 4075.2432,
      "high": 4277.36295
    }
  },
  "not_invoice": true
}