Berkeley Function Calling Leaderboard v3

A function-calling benchmark for tool selection, schema adherence, and argument correctness, covering single-turn, parallel, irrelevance and multi-turn subsets.

5 models

Higher is better.

1Ternary Bonsai 2 27BPrism ML74.9
2Qwen3-Coder 480B-A35BAlibaba68.7
3Qwen3 235B-A22BAlibaba68.0
4QwQ-32BAlibaba66.4
5Qwen3 32BAlibaba63.0

Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric: Tool invocation and schema evaluation
Tasks: Function-calling tasks
Authors: Shishir G. Patil et al. (2025)
Source: The Berkeley Function Calling Leaderboard (BFCL): From Tool Use to Agentic Evaluation of Large Language Models ↗

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.