{ "cost_weights_assumed": { "small": 1, "medium": 3, "large": 10, "reasoning": 15 }, "sets": { "heldout_combo": { "n": 1200, "accuracy": 0.9841666666666666, "under_routing_rate": 0.0016666666666666668, "over_routing_rate": 0.014166666666666666, "cost_ratio_vs_gold": 1.0072413793103447, "overspend_rate": 0.009655172413793104, "confusion_matrix": { "rows_gold_cols_pred": [ "small", "medium", "large", "reasoning" ], "matrix": [ [ 293, 7, 0, 0 ], [ 0, 290, 10, 0 ], [ 0, 1, 299, 0 ], [ 1, 0, 0, 299 ] ] }, "per_tier_accuracy": { "small": 0.9766666666666667, "medium": 0.9666666666666667, "large": 0.9966666666666667, "reasoning": 0.9966666666666667 }, "per_language_accuracy": { "en": 0.9916666666666667, "ko": 0.9966666666666667, "vi": 0.9833333333333333, "vi_nodiac": 0.9666666666666667 }, "per_family": { "bash_script": { "accuracy": 0.9803921568627451, "n": 51, "predicted": { "medium": 50, "large": 1 } }, "cicd_monorepo": { "accuracy": 1.0, "n": 52, "predicted": { "large": 52 } }, "cipher": { "accuracy": 1.0, "n": 64, "predicted": { "reasoning": 64 } }, "constraint_puzzle": { "accuracy": 1.0, "n": 74, "predicted": { "reasoning": 74 } }, "csv_to_json": { "accuracy": 0.9857142857142858, "n": 70, "predicted": { "small": 69, "medium": 1 } }, "dockerfile": { "accuracy": 0.9310344827586207, "n": 58, "predicted": { "medium": 54, "large": 4 } }, "explain_snippet": { "accuracy": 1.0, "n": 78, "predicted": { "medium": 78 } }, "graph_proof": { "accuracy": 1.0, "n": 53, "predicted": { "reasoning": 53 } }, "multi_transcript": { "accuracy": 1.0, "n": 61, "predicted": { "large": 61 } }, "paradox": { "accuracy": 0.9830508474576272, "n": 59, "predicted": { "reasoning": 58, "small": 1 } }, "physics_derive": { "accuracy": 1.0, "n": 50, "predicted": { "reasoning": 50 } }, "prd": { "accuracy": 0.984375, "n": 64, "predicted": { "large": 63, "medium": 1 } }, "regex": { "accuracy": 1.0, "n": 47, "predicted": { "medium": 47 } }, "repo_qa": { "accuracy": 1.0, "n": 65, "predicted": { "large": 65 } }, "study_plan": { "accuracy": 0.9242424242424242, "n": 66, "predicted": { "medium": 61, "large": 5 } }, "synonym": { "accuracy": 1.0, "n": 54, "predicted": { "small": 54 } }, "timezone": { "accuracy": 1.0, "n": 63, "predicted": { "small": 63 } }, "vendor_compare": { "accuracy": 1.0, "n": 58, "predicted": { "large": 58 } }, "word_count": { "accuracy": 0.9629629629629629, "n": 54, "predicted": { "small": 52, "medium": 2 } }, "yes_no_fact": { "accuracy": 0.9322033898305084, "n": 59, "predicted": { "small": 55, "medium": 4 } } }, "pass": true }, "heldout_style": { "n": 1200, "accuracy": 0.96, "under_routing_rate": 0.0016666666666666668, "over_routing_rate": 0.03833333333333333, "cost_ratio_vs_gold": 1.0241379310344827, "overspend_rate": 0.02517241379310345, "confusion_matrix": { "rows_gold_cols_pred": [ "small", "medium", "large", "reasoning" ], "matrix": [ [ 261, 32, 0, 7 ], [ 1, 293, 4, 2 ], [ 0, 1, 298, 1 ], [ 0, 0, 0, 300 ] ] }, "per_tier_accuracy": { "small": 0.87, "medium": 0.9766666666666667, "large": 0.9933333333333333, "reasoning": 1.0 }, "per_language_accuracy": { "en": 0.9791666666666666, "ko": 0.95, "vi": 0.9611111111111111, "vi_nodiac": 0.9533333333333334 }, "per_family": { "bash_script": { "accuracy": 1.0, "n": 58, "predicted": { "medium": 58 } }, "cicd_monorepo": { "accuracy": 0.9583333333333334, "n": 48, "predicted": { "large": 46, "reasoning": 1, "medium": 1 } }, "cipher": { "accuracy": 1.0, "n": 62, "predicted": { "reasoning": 62 } }, "constraint_puzzle": { "accuracy": 1.0, "n": 72, "predicted": { "reasoning": 72 } }, "csv_to_json": { "accuracy": 0.9661016949152542, "n": 59, "predicted": { "small": 57, "medium": 2 } }, "dockerfile": { "accuracy": 0.9636363636363636, "n": 55, "predicted": { "medium": 53, "reasoning": 1, "large": 1 } }, "explain_snippet": { "accuracy": 0.9714285714285714, "n": 70, "predicted": { "medium": 68, "small": 1, "reasoning": 1 } }, "graph_proof": { "accuracy": 1.0, "n": 47, "predicted": { "reasoning": 47 } }, "multi_transcript": { "accuracy": 1.0, "n": 78, "predicted": { "large": 78 } }, "paradox": { "accuracy": 1.0, "n": 60, "predicted": { "reasoning": 60 } }, "physics_derive": { "accuracy": 1.0, "n": 59, "predicted": { "reasoning": 59 } }, "prd": { "accuracy": 1.0, "n": 57, "predicted": { "large": 57 } }, "regex": { "accuracy": 1.0, "n": 56, "predicted": { "medium": 56 } }, "repo_qa": { "accuracy": 1.0, "n": 60, "predicted": { "large": 60 } }, "study_plan": { "accuracy": 0.9508196721311475, "n": 61, "predicted": { "medium": 58, "large": 3 } }, "synonym": { "accuracy": 0.86, "n": 50, "predicted": { "medium": 6, "small": 43, "reasoning": 1 } }, "timezone": { "accuracy": 0.9298245614035088, "n": 57, "predicted": { "small": 53, "medium": 1, "reasoning": 3 } }, "vendor_compare": { "accuracy": 1.0, "n": 57, "predicted": { "large": 57 } }, "word_count": { "accuracy": 0.8333333333333334, "n": 66, "predicted": { "small": 55, "medium": 10, "reasoning": 1 } }, "yes_no_fact": { "accuracy": 0.7794117647058824, "n": 68, "predicted": { "small": 53, "medium": 13, "reasoning": 2 } } }, "pass": true } } }