{"$k":["id","title","columns","rows"],"$r":[["thinking-bill-hard-cells","Reasoning tokens and list-price cost per call, hard tasks (calculation)",{"$k":["key","label","unit"],"$r":[["config","Configuration","text"],["calls","Calls","count"],["strict","Strict passes","text"],["medOut","Median output tokens","tokens"],["medRea","Median reasoning tokens","tokens"],["tokenRange","Output / reasoning tokens, lowest to highest call","text"],["costRange","Reasoning / total cost, lowest to highest call (USD, calculation)","text"],["costShareRange","Reasoning cost share, lowest to highest call (calculation)","text"],["zero","Calls with 0 reasoning tokens","count"],["medShare","Median call: reasoning share of output","rate"],["shareRange","Share, lowest to highest call","text"],["pooled","Pooled share (all reasoning ÷ all output)","rate"],["meanReasoning","Reasoning cost per call (USD, calculation)","usd"],["meanVisible","Remaining output cost per call (USD, calculation)","usd"],["meanInput","Input cost per call (USD)","usd"],["meanTotal","Total cost per call (USD)","usd"],["costShare","Pooled reasoning share of list-price cost (calculation)","rate"],["reasoningPerPass","Reasoning cost per strict pass (USD)","usd"],["totalPerPass","Total cost per strict pass (USD)","usd"]]},{"$k":["config","calls","strict","medOut","medRea","tokenRange","costRange","costShareRange","zero","medShare","shareRange","pooled","meanReasoning","meanVisible","meanInput","meanTotal","costShare","reasoningPerPass","totalPerPass"],"$r":[["Claude Haiku 4.5 · Claude Code",24,"11/24 (95% Wilson 28% to 65%)",5064,4556,"1899 to 9321 / 1452 to 8569","$0.0073 to $0.0428 / $0.0148 to $0.0505","40% to 90%",0,0.9168,"76% to 99%",0.9317,0.024492,0.001795,0.00451,0.030798,0.7953,0.053438,0.067196],["Claude Sonnet 5.5 · Claude Code",24,"24/24 (95% Wilson 86% to 100%)",1050,585,"176 to 3895 / 0 to 3060","$0.0000 to $0.0306 / $0.0051 to $0.0424","0% to 74%",7,0.5454,"0% to 96%",0.6448,0.006665,0.003672,0.004012,0.014349,0.4645,0.006665,0.014349],["Claude Opus 5.5 · Claude Code",24,"24/24 (95% Wilson 86% to 100%)",945,529,"323 to 2531 / 100 to 1778","$0.0020 to $0.0356 / $0.0131 to $0.0572","10% to 74%",0,0.5479,"30% to 96%",0.6102,0.012528,0.008003,0.00771,0.028241,0.4436,0.012528,0.028241],["Claude Opus 5.5 (high) · Claude Code",24,"24/24 (95% Wilson 86% to 100%)",1052,614,"285 to 4052 / 103 to 3301","$0.0021 to $0.0660 / $0.0121 to $0.0877","17% to 77%",0,0.5443,"36% to 96%",0.6922,0.017969,0.00799,0.007407,0.033366,0.5385,0.017969,0.033366],["Claude Fable 5.1 · Claude Code",24,"24/24 (95% Wilson 86% to 100%)",1366,889,"318 to 6465 / 75 to 5889","$0.0037 to $0.2944 / $0.0322 to $0.3399","5% to 87%",0,0.6424,"23% to 97%",0.7417,0.053696,0.018702,0.02091,0.093308,0.5755,0.053696,0.093308],["GPT-6.1 Sol (medium) · Codex CLI",16,"16/16 (95% Wilson 81% to 100%)",335,150,"237 to 1766 / 61 to 839","$0.0006 to $0.0084 / $0.0099 to $0.0422","2% to 20%",0,0.4633,"11% to 87%",0.429,0.002273,0.003025,0.020339,0.025637,0.0887,0.002273,0.025637],["GPT-6.1 Sol (high) · Codex CLI",16,"16/16 (95% Wilson 81% to 100%)",436,225,"284 to 2569 / 144 to 1750","$0.0014 to $0.0175 / $0.0092 to $0.0307","8% to 60%",0,0.5701,"29% to 91%",0.5861,0.004114,0.002905,0.008117,0.015137,0.2718,0.004114,0.015137]]}],["thinking-bill-effort-cells","Reasoning tokens and list-price cost by effort, every effort-ladder cell (calculation)",{"$k":["key","label","unit"],"$r":[["config","Configuration","text"],["calls","Calls","count"],["strict","Strict passes","text"],["medOut","Median output tokens","tokens"],["medRea","Median reasoning tokens","tokens"],["tokenRange","Output / reasoning tokens, lowest to highest call","text"],["costRange","Reasoning / total cost, lowest to highest call (USD, calculation)","text"],["costShareRange","Reasoning cost share, lowest to highest call (calculation)","text"],["zero","Calls with 0 reasoning tokens","count"],["medShare","Median call: reasoning share of output","rate"],["shareRange","Share, lowest to highest call","text"],["pooled","Pooled share (all reasoning ÷ all output)","rate"],["meanReasoning","Reasoning cost per call (USD, calculation)","usd"],["meanVisible","Remaining output cost per call (USD, calculation)","usd"],["meanInput","Input cost per call (USD)","usd"],["meanTotal","Total cost per call (USD)","usd"],["costShare","Pooled reasoning share of list-price cost (calculation)","rate"],["reasoningPerPass","Reasoning cost per strict pass (USD)","usd"],["totalPerPass","Total cost per strict pass (USD)","usd"]]},{"$k":["config","calls","strict","medOut","medRea","tokenRange","costRange","costShareRange","zero","medShare","shareRange","pooled","meanReasoning","meanVisible","meanInput","meanTotal","costShare","reasoningPerPass","totalPerPass"],"$r":[["Claude Sonnet 5.5 (low) · Claude Code",16,"16/16 (95% Wilson 81% to 100%)",667,273,"176 to 2263 / 0 to 1489","$0.0000 to $0.0149 / $0.0051 to $0.0261","0% to 71%",8,0.2142,"0% to 95%",0.5327,0.004317,0.003787,0.004088,0.012191,0.3541,0.004317,0.012191],["Claude Sonnet 5.5 (medium) · Claude Code",16,"16/16 (95% Wilson 81% to 100%)",770,422,"224 to 2836 / 0 to 2093","$0.0000 to $0.0209 / $0.0056 to $0.0318","0% to 75%",5,0.5162,"0% to 96%",0.6158,0.005946,0.003709,0.003865,0.01352,0.4398,0.005946,0.01352],["Claude Sonnet 5.5 (high) · Claude Code",16,"16/16 (95% Wilson 81% to 100%)",1192,745,"220 to 4187 / 0 to 3610","$0.0000 to $0.0361 / $0.0056 to $0.0453","0% to 80%",2,0.6131,"0% to 96%",0.7297,0.009369,0.003471,0.003866,0.016705,0.5608,0.009369,0.016705],["Claude Sonnet 5.5 · Claude Code",16,"16/16 (95% Wilson 81% to 100%)",1054,668,"176 to 2185 / 0 to 1614","$0.0000 to $0.0161 / $0.0051 to $0.0253","0% to 74%",4,0.5601,"0% to 96%",0.6368,0.006299,0.003593,0.004086,0.013978,0.4507,0.006299,0.013978],["Claude Opus 5.5 (low) · Claude Code",16,"16/16 (95% Wilson 81% to 100%)",594,87,"220 to 1569 / 0 to 791","$0.0000 to $0.0158 / $0.0108 to $0.0380","0% to 67%",7,0.3065,"0% to 94%",0.3785,0.005031,0.008261,0.00786,0.021152,0.2379,0.005031,0.021152],["Claude Opus 5.5 (medium) · Claude Code",16,"16/16 (95% Wilson 81% to 100%)",853,518,"301 to 3075 / 109 to 1993","$0.0022 to $0.0399 / $0.0125 to $0.0681","16% to 74%",0,0.5353,"28% to 96%",0.609,0.01344,0.00863,0.007405,0.029475,0.456,0.01344,0.029475],["Claude Opus 5.5 (high) · Claude Code",16,"16/16 (95% Wilson 81% to 100%)",1052,614,"285 to 4052 / 103 to 3301","$0.0021 to $0.0660 / $0.0121 to $0.0877","17% to 77%",0,0.5369,"36% to 96%",0.6865,0.018034,0.008236,0.007407,0.033677,0.5355,0.018034,0.033677],["Claude Opus 5.5 · Claude Code",16,"16/16 (95% Wilson 81% to 100%)",945,538,"323 to 2531 / 100 to 1778","$0.0020 to $0.0356 / $0.0131 to $0.0572","10% to 72%",0,0.5435,"31% to 95%",0.6221,0.013104,0.007961,0.00786,0.028925,0.453,0.013104,0.028925],["GPT-6.1 Sol (low) · Codex CLI",16,"16/16 (95% Wilson 81% to 100%)",284,63,"172 to 1310 / 0 to 458","$0.0000 to $0.0046 / $0.0092 to $0.0264","0% to 22%",4,0.2351,"0% to 84%",0.2909,0.001223,0.00298,0.008635,0.012837,0.0952,0.001223,0.012837],["GPT-6.1 Sol (medium) · Codex CLI",16,"16/16 (95% Wilson 81% to 100%)",335,150,"237 to 1766 / 61 to 839","$0.0006 to $0.0084 / $0.0099 to $0.0422","2% to 20%",0,0.4633,"11% to 87%",0.429,0.002273,0.003025,0.020339,0.025637,0.0887,0.002273,0.025637],["GPT-6.1 Sol (high) · Codex CLI",16,"16/16 (95% Wilson 81% to 100%)",436,225,"284 to 2569 / 144 to 1750","$0.0014 to $0.0175 / $0.0092 to $0.0307","8% to 60%",0,0.5701,"29% to 91%",0.5861,0.004114,0.002905,0.008117,0.015137,0.2718,0.004114,0.015137]]}],["thinking-bill-short-cells","Reasoning tokens and list-price cost per call, five short tasks (calculation)",{"$k":["key","label","unit"],"$r":[["config","Configuration","text"],["calls","Calls","count"],["strict","Strict passes","text"],["medOut","Median output tokens","tokens"],["medRea","Median reasoning tokens","tokens"],["tokenRange","Output / reasoning tokens, lowest to highest call","text"],["costRange","Reasoning / total cost, lowest to highest call (USD, calculation)","text"],["costShareRange","Reasoning cost share, lowest to highest call (calculation)","text"],["zero","Calls with 0 reasoning tokens","count"],["medShare","Median call: reasoning share of output","rate"],["shareRange","Share, lowest to highest call","text"],["pooled","Pooled share (all reasoning ÷ all output)","rate"],["meanReasoning","Reasoning cost per call (USD, calculation)","usd"],["meanVisible","Remaining output cost per call (USD, calculation)","usd"],["meanInput","Input cost per call (USD)","usd"],["meanTotal","Total cost per call (USD)","usd"],["costShare","Pooled reasoning share of list-price cost (calculation)","rate"],["reasoningPerPass","Reasoning cost per strict pass (USD)","usd"],["totalPerPass","Total cost per strict pass (USD)","usd"]]},{"$k":["config","calls","strict","medOut","medRea","tokenRange","costRange","costShareRange","zero","medShare","shareRange","pooled","meanReasoning","meanVisible","meanInput","meanTotal","costShare","reasoningPerPass","totalPerPass"],"$r":[["Claude Haiku 4.5 · Claude Code",15,"15/15 (95% Wilson 80% to 100%)",367,297,"275 to 2851 / 212 to 2643","$0.0011 to $0.0132 / $0.0051 to $0.0180","20% to 74%",0,0.9019,"73% to 98%",0.9049,0.004133,0.000434,0.00379,0.008358,0.4945,0.004133,0.008358],["Claude Sonnet 5.5 · Claude Code",15,"12/15 (95% Wilson 55% to 93%)",107,0,"44 to 745 / 0 to 542","$0.0000 to $0.0054 / $0.0034 to $0.0102","0% to 53%",12,0,"0% to 73%",0.4468,0.000882,0.001092,0.003015,0.004989,0.1768,0.001103,0.006236],["Claude Opus 5.5 · Claude Code",15,"15/15 (95% Wilson 80% to 100%)",64,0,"44 to 853 / 0 to 616","$0.0000 to $0.0123 / $0.0059 to $0.0223","0% to 55%",9,0,"0% to 93%",0.5912,0.002585,0.001788,0.005714,0.010088,0.2563,0.002585,0.010088],["Claude Opus 5.5 (low) · Claude Code",15,"15/15 (95% Wilson 80% to 100%)",64,0,"44 to 637 / 0 to 405","$0.0000 to $0.0081 / $0.0058 to $0.0179","0% to 45%",9,0,"0% to 93%",0.4098,0.001253,0.001805,0.005232,0.00829,0.1512,0.001253,0.00829],["Claude Opus 5.5 (high) · Claude Code",15,"15/15 (95% Wilson 80% to 100%)",78,34,"44 to 1094 / 0 to 857","$0.0000 to $0.0171 / $0.0059 to $0.0271","0% to 63%",7,0.4359,"0% to 93%",0.6562,0.003448,0.001807,0.005233,0.010488,0.3288,0.003448,0.010488],["Claude Fable 5.1 · Claude Code",15,"15/15 (95% Wilson 80% to 100%)",64,0,"4 to 908 / 0 to 672","$0.0000 to $0.0336 / $0.0049 to $0.0584","0% to 67%",12,0,"0% to 74%",0.5724,0.005957,0.00445,0.010133,0.020539,0.29,0.005957,0.020539],["GPT-6.1 Sol (low) · Codex CLI",10,"10/10 (95% Wilson 72% to 100%)",42,20,"28 to 225 / 0 to 103","$0.0000 to $0.0010 / $0.0074 to $0.0265","0% to 11%",4,0.2528,"0% to 71%",0.3132,0.000332,0.000728,0.008924,0.009984,0.0333,0.000332,0.009984],["GPT-6.1 Sol (medium) · Codex CLI",15,"15/15 (95% Wilson 80% to 100%)",42,20,"27 to 295 / 0 to 137","$0.0000 to $0.0014 / $0.0054 to $0.0269","0% to 23%",6,0.4105,"0% to 71%",0.4142,0.000512,0.000724,0.014404,0.01564,0.0327,0.000512,0.01564],["GPT-6.1 Sol (high) · Codex CLI",15,"15/15 (95% Wilson 80% to 100%)",42,21,"28 to 457 / 0 to 289","$0.0000 to $0.0029 / $0.0066 to $0.0281","0% to 29%",6,0.5806,"0% to 76%",0.5814,0.001007,0.000725,0.011484,0.013216,0.0762,0.001007,0.013216]]}],["thinking-bill-time-link","Does reasoning track time? Rank correlation and slope per model (calculation)",{"$k":["key","label","unit"],"$r":[["model","Model and route","text"],["calls","Calls","count"],["rho","Spearman: reasoning tokens vs time","score"],["rhoCi","Leave-one-task-out range (not a 95% interval)","text"],["rhoOut","Spearman: output tokens vs time","score"],["perK","Seconds per 1,000 reasoning tokens (all calls)","seconds"],["withinK","Seconds per 1,000 reasoning tokens (within task)","seconds"],["withinCi","Within-task slope, leave-one-task-out range","text"],["medRea","Median reasoning tokens","tokens"],["medTime","Median total time (s)","seconds"],["timeRange","Total time, lowest to highest call (s; not an interval)","text"]]},{"$k":["model","calls","rho","rhoCi","rhoOut","perK","withinK","withinCi","medRea","medTime","timeRange"],"$r":[["Claude Haiku 4.5 · Claude Code",24,0.982,"0.98 to 0.99",0.986,8.408,8.817,"8.6 to 9.1",4556,39.01,"15.27 to 75.13"],["Claude Sonnet 5.5 · Claude Code",72,0.921,"0.88 to 0.94",0.94,9.28,7.649,"7.4 to 7.8",586,7.75,"2.26 to 35.81"],["Claude Opus 5.5 · Claude Code",80,0.852,"0.82 to 0.89",0.97,13.091,11.754,"5.7 to 12.7",511,9.01,"3.34 to 63.00"],["Claude Fable 5.1 · Claude Code",24,0.924,"0.89 to 0.93",0.941,14.39,14.454,"14.4 to 22.6",889,16.13,"4.46 to 90.00"],["GPT-6.1 Sol · Codex CLI",48,0.556,"0.34 to 0.65",0.85,50.374,36.351,"31.8 to 36.7",189,14.51,"7.94 to 92.20"]]}],["thinking-bill-sum-check","Consistency check for reasoning within output tokens (calculation)",{"$k":["key","label","unit"],"$r":[["route","CLI route","text"],["calls","Calls","count"],["withReasoning","Calls with more than 0 reasoning tokens","count"],["zero","Calls that reported 0 (kept as 0)","count"],["over","Calls with reasoning above output","count"],["slope","Output tokens gained per extra reasoning token (within task)","ratio"],["slopeCi","Leave-one-task-out range (not a 95% interval)","text"],["cpvisible","Characters per token of output minus reasoning","score"],["cpoutZero","Characters per output token, calls with 0 reasoning","score"],["cpoutWith","Characters per output token, calls with reasoning","score"],["verdict","Reasoning is part of output","text"]]},[{"route":"Claude Code","calls":290,"withReasoning":212,"zero":78,"over":0,"slope":0.993,"slopeCi":"0.982 to 1.009","cpvisible":1.94,"cpoutZero":1.98,"cpoutWith":0.45,"verdict":"consistent with inclusion; not proof"},{"route":"Codex CLI","calls":88,"withReasoning":68,"zero":20,"over":0,"slope":0.967,"slopeCi":"0.965 to 0.986","cpvisible":2.99,"cpoutZero":2.76,"cpoutWith":1.36,"verdict":"consistent with inclusion; not proof"}]]]}