Skip to content

Benchmark

BFCL Non-Live AST

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.

Models measured
49
Scores recorded
76
Sources
1
IQ weight
1.15

Score distribution

77.5%
Mean
84%
Median
17.4%
Std dev
59.8%
10th percentile
88.8%
90th percentile
0%
Min
0%90.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.15 for chat models, measured over 49 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Gemini 3 Pro PreviewPrompt· Berkeley Function-Calling Leaderboard data90.7%
2Qwen3 235B A22B Instruct 2507Prompt· Berkeley Function-Calling Leaderboard data90.3%
3Qwen3 32BPrompt· Berkeley Function-Calling Leaderboard data90.3%
4Mistral SmallPrompt· Berkeley Function-Calling Leaderboard data89.7%
5Claude Opus 4.5Prompt· Berkeley Function-Calling Leaderboard data89.7%
6Qwen3 14BPrompt· Berkeley Function-Calling Leaderboard data89.5%
7Llama 4 Scoutfc· Berkeley Function-Calling Leaderboard data89.4%
8Qwen3 30B A3B InstructPrompt· Berkeley Function-Calling Leaderboard data88.9%
9Qwen3 32Bfc· Berkeley Function-Calling Leaderboard data88.8%
10GPT 4.1Prompt· Berkeley Function-Calling Leaderboard data88.7%
11Claude Sonnet 4.5fc· Berkeley Function-Calling Leaderboard data88.7%
12Llama 4 Maverickfc· Berkeley Function-Calling Leaderboard data88.7%
13Claude Opus 4.5fc· Berkeley Function-Calling Leaderboard data88.6%
14Qwen3 8BPrompt· Berkeley Function-Calling Leaderboard data88.6%
15Open Mistral NemoPrompt· Berkeley Function-Calling Leaderboard data88.5%
16Grok 4.1 Fastfc· Berkeley Function-Calling Leaderboard data88.3%
17Gemini 2.5 FlashPrompt· Berkeley Function-Calling Leaderboard data88.1%
18Llama 3.3 70B Instructfc· Berkeley Function-Calling Leaderboard data88%
19Qwen3 4B Instruct 2507fc· Berkeley Function-Calling Leaderboard data87.9%
20Qwen3 8Bfc· Berkeley Function-Calling Leaderboard data87.6%
21GLM 4.6FC thinking· Berkeley Function-Calling Leaderboard data87.6%
22Gemma 3 27B ITPrompt· Berkeley Function-Calling Leaderboard data87.2%
23Nova 2 Litefc· Berkeley Function-Calling Leaderboard data87%
24Gemini 2.5 Flash-Litefc· Berkeley Function-Calling Leaderboard data86.6%
25Nova Profc· Berkeley Function-Calling Leaderboard data86.6%
26Claude Haiku 4.5fc· Berkeley Function-Calling Leaderboard data86.5%
27Qwen3 4B Instruct 2507Prompt· Berkeley Function-Calling Leaderboard data86.4%
28Command Afc· Berkeley Function-Calling Leaderboard data86.3%
29Qwen3 30B A3B Instructfc· Berkeley Function-Calling Leaderboard data85.8%
30Gemini 3 Pro Previewfc· Berkeley Function-Calling Leaderboard data85.8%
31DeepSeek V3.2 ExpPrompt + Thinking· Berkeley Function-Calling Leaderboard data85.5%
32Grok 4fc· Berkeley Function-Calling Leaderboard data85.4%
33Mistral Medium 3· Berkeley Function-Calling Leaderboard data85.3%
34Gemini 2.5 Flashfc· Berkeley Function-Calling Leaderboard data85%
35Qwen3 14Bfc· Berkeley Function-Calling Leaderboard data84.9%
36Mistral Largefc· Berkeley Function-Calling Leaderboard data84.7%
37GPT-4.1 miniPrompt· Berkeley Function-Calling Leaderboard data84.6%
38Llama 3.1 8B InstructPrompt· Berkeley Function-Calling Leaderboard data84%
39Gemini 2.5 Flash-LitePrompt· Berkeley Function-Calling Leaderboard data83.9%
40GPT-4.1 minifc· Berkeley Function-Calling Leaderboard data83.8%
41Mistral LargePrompt· Berkeley Function-Calling Leaderboard data83%
42Qwen3 1.7Bfc· Berkeley Function-Calling Leaderboard data82.9%
43Open Mistral Nemofc· Berkeley Function-Calling Leaderboard data82.8%
44GPT 4.1fc· Berkeley Function-Calling Leaderboard data82.8%
45Grok 4Prompt· Berkeley Function-Calling Leaderboard data82.8%
46Llama 3.2 3Bfc· Berkeley Function-Calling Leaderboard data82.7%
47OpenAI o3Prompt· Berkeley Function-Calling Leaderboard data81.9%
48GPT 5.2fc· Berkeley Function-Calling Leaderboard data81.8%
49Kimi K2 Instructfc· Berkeley Function-Calling Leaderboard data81.6%
50o4 MiniPrompt· Berkeley Function-Calling Leaderboard data81.3%

Showing 1–50 of 76 results.

Data credits: Berkeley Function-Calling Leaderboard (Gorilla, UC Berkeley), Apache-2.0 (licence)