Share your thoughts, 1 month free Claude Pro on usSee more

Function Calling on BFCL (Accuracy %)

86.67Accuracy (%)

Opus-4.6 (1M)

Updated 1d ago

Evaluation Results

Method	Links
Opus-4.6 (1M) 2026.05		86.67
Official Instruct Model 2026.05		84
GLM-4.7 & ANDES (Ours) 2026.05		78.67
Opus-4.6 2026.05		75.92
GLM-4.7 (Scaffold-only) 2026.05		70
Opus-4.7 (xHigh) 2026.05		62.33
GPT-5.2 2026.05		33.33
GPT-5.4 (High) 2026.05		29.67
Gemini-3.1-Pro 2026.05		27.67
Base Model (SmolLM3-3B) 2026.05		0
Sonnet-4.5 2026.05		0
Qwen3-Max 2026.05		0
Kimi-K2-Thinking 2026.05		0
MiniMax-M2.1 2026.05		0
GPT-5.1-Codex-Max 2026.05		0
MiniMax-M2.5 2026.05		0
GLM-4.7 (OpenCode) 2026.05		0