Calibration
MCP-Tasks → τ²-bench results
τ²-bench results is estimated from MCP-Tasks with a offset logistic curve fitted on 5 models measured on both: y = 0.9706 + (0.8930 − 0.9706) / (1 + exp(−124.17·(x − 0.6589))), R² = 0.39, cross-validated error 6.9 pp. It is used for 0 estimates.