ResearchBenchmarksTool-calling on BFCL Generalization (test)Follow87.5Final AccuracyMAC84.3885.198686.81Mar 16, 2026Evaluation ResultsMethodMethodLinksFinal AccuracyImprovementMACbaseline_accuracy=82%baseline_accuracy=82%2026.0387.55.5GEPAbaseline_accuracy=82%baseline_accuracy=82%2026.03864MIPRObaseline_accuracy=82%baseline_accuracy=82%2026.0384.52.5