arm,committee,aggregation,cost_usd_mean,latency_s_mean,tokens_mean,cost_usd_total,mean_abs_log_error,n A_hedonic,none,model,0.0,1.2269304174544202e-06,0.0,0.0,0.19056946791554397,96935.0 A_hedonic_spatial,none,model,0.0,0.0,0.0,0.0,0.20801062571921,8988.0 B_hierarchical,none,model,0.0,9.187243098888168e-07,0.0,0.0,0.18187600949116897,96935.0 C_catboost,none,model,0.0,1.196263392896917e-06,0.0,0.0,0.1767604987174085,96935.0 C_gbm,none,model,0.0,2.9650431629751662e-06,0.0,0.0,0.1792328866925307,96935.0 C_gbm_spatial,none,model,0.0,0.0,0.0,0.0,0.1763518946028651,8988.0 C_xgboost,none,model,0.0,9.557852377807515e-07,0.0,0.0,0.18017295155532076,96935.0 D_avm,none,model,0.0,0.00022600315588784827,0.0,0.0,0.17147330557935556,96935.0 D_avm_spatial,none,model,0.0,0.0,0.0,0.0,0.16067798066287955,8988.0 E_single_llm,mixed_random,median,0.0218719625,19.7346875,6596.4125,1.749757,0.1655996928706865,80.0 E_single_llm_small,mixed_random,median,0.0017041766666666665,19.08168,8132.62,0.511253,0.18203351780088417,300.0 F_independent,mixed_random,median,0.23228025,79.1258125,73947.1625,18.58242,0.15871028571985762,80.0 F_independent_small,mixed_random,median,0.01875933,78.47935666666667,89503.81333333334,5.627799,0.1760399627264085,300.0 G_debate_r2,mixed_random,median,1.521668475,336.19342500000005,521126.9375,121.733478,0.1575785832187693,80.0 G_debate_r2_small,mixed_random,median,0.12528112,358.38188333333335,667794.92,37.584336,0.1802127691886514,300.0 H_debate_skeptic,mixed_random,median,1.3501288875,304.8450875,459056.7875,108.010311,0.15764573155979728,80.0 I_placebo,mixed_random,median,1.3973111875000002,268.12165,491998.425,111.784895,0.15739760350653925,80.0 J_cost_matched,mixed_random,median,0.25540555,93.8318125,91937.3875,20.432444,0.1633342294817131,80.0