List your company

Agentic Reasoning on BFCL v4 (non-live and live)

86.45AccuracyCopT
77.173279.581681.9984.3984May 19, 2026

Evaluation results

MethodLinks
CopTBackbone=Qwen2.5-35B-A3B, Reasoning Effort=Higher2026.0586.45168
CopTBackbone=Qwen2.5-35B-A3B, Reasoning Effort=Standard2026.0586.17130
CoTBackbone=Qwen2.5-35B-A3B2026.0585.77235
CopTBackbone=Qwen2.5-2B, Reasoning Effort=Higher2026.0578.37164
CopTBackbone=Qwen2.5-2B, Reasoning Effort=Standard2026.0578.01139
CoTBackbone=Qwen2.5-2B2026.0577.53234