Total Cost ($)

BFCL · 50 models

Rank	Model	Score
1	Qwen3-0.6B (FC)	0.46
2	Llama-3.1-Nemotron-Ultra-253B-v1 (FC)	0.72
3	Qwen3-14B (Prompt)	1.35
4	Granite-4.0-350m (FC)	1.44
5	Command R7B (FC)	1.5
6	Qwen3-30B-A3B-Instruct-2507 (Prompt)	1.56
7	Llama-3.2-1B-Instruct (FC)	1.64
8	Falcon3-1B-Instruct (FC)	1.72
9	Amazon-Nova-Micro-v1:0 (FC)	1.81
10	Arch-Agent-1.5B	2.45
11	Qwen3-235B-A22B-Instruct-2507 (FC)	2.5
12	xLAM-2-1b-fc-r (FC)	2.79
13	Hammer2.1-0.5b (FC)	2.82
14	Command A Reasoning (FC)	3.04
15	Qwen3-235B-A22B-Instruct-2507 (Prompt)	3.12
16	xLAM-2-3b-fc-r (FC)	3.36
17	Qwen3-14B (FC)	3.38
18	Gemma-3-1b-it (Prompt)	3.4
19	Qwen3-0.6B (Prompt)	3.65
20	Arch-Agent-3B	3.7
21	Gemma-3-4b-it (Prompt)	4.14
22	Qwen3-1.7B (FC)	4.33
23	GLM-4.6 (FC thinking)	4.64
24	Hammer2.1-7b (FC)	4.99
25	Mistral-small-2506 (FC)	5.2
26	Granite-20b-FunctionCalling (FC)	5.23
27	Qwen3-30B-A3B-Instruct-2507 (FC)	5.62
28	GPT-4.1-nano-2025-04-14 (FC)	5.66
29	xLAM-2-32b-fc-r (FC)	6
30	Moonshotai-Kimi-K2-Instruct (FC)	6.19
31	Llama-3.2-3B-Instruct (FC)	6.2
32	Qwen3-4B-Instruct-2507 (FC)	6.37
33	DeepSeek-V3.2-Exp (FC)	6.71
34	Hammer2.1-1.5b (FC)	6.83
35	Mistral-Small-2506 (Prompt)	6.91
36	Gemini-2.5-Flash-Lite (Prompt)	7.05
37	GPT-4.1-nano-2025-04-14 (Prompt)	7.42
38	Llama-3.1-8B-Instruct (Prompt)	7.49
39	Gemini-2.5-Flash-Lite (FC)	7.55
40	Open-Mistral-Nemo-2407 (FC)	8.12
41	Phi-4 (Prompt)	8.72
42	GPT-5-nano-2025-08-07 (FC)	8.79
43	Arch-Agent-32B	8.87
44	Granite-3.1-8B-Instruct (FC)	9.32
45	Gemma-3-12b-it (Prompt)	10.77
46	Hammer2.1-3b (FC)	10.89
47	Gemma-3-27b-it (Prompt)	11.82
48	RZN-T (Prompt)	12.31
49	Open-Mistral-Nemo-2407 (Prompt)	13.8
50	Nanbeige4-3B-Thinking-2511 (FC)	14.14