Benchmarks / Berkeley Function Calling Leaderboard (BFCL) V4

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

Berkeley Function Calling Leaderboard (BFCL) V4

Declining to call a function when none of those offered fits.

Data older than 90 days: last updated 16 Dec 2025

Results dated
16 Dec 2025
Results
109 configurations of 81 models
Unit
% correct
Licence
Apache License 2.0

Berkeley Function Calling Leaderboard (BFCL) V4 in brief

What it measures. Berkeley Function Calling Leaderboard (BFCL) V4 is a third-party benchmark, shown here with credit. Measured here: Declining to call a function when none of those offered fits.

Latest results. , 109 configurations ranked.

Top 3 on Irrelevance detection (% correct, higher is better):

  1. Ministral 8B (2410) (Mistral AI): 100.0%
  2. Llama 3.1 Nemotron Ultra 253B v1 (NVIDIA): 100.0%
  3. BitAgent Bounty 8B (Bittensor): 97.5%

The first 2 are too close to tell apart, so read the order among them as a tie.

How to cite. Berkeley Function Calling Leaderboard (BFCL) V4, via Spring Prompt, results dated 16 December 2025. https://springprompt.com/benchmarks/bfcl These results belong to Berkeley Function Calling Leaderboard (BFCL) V4, not to us. Licence: Apache License 2.0. Credit Berkeley Function Calling Leaderboard (BFCL) V4 first.

Irrelevance detection: Grok 4.1 Fast (reasoning)

Top 15 of 109 results · % correct, higher is better · ≈ cannot be told apart from the leader. Choose a model to highlight it.Clear highlight

  1. 1≈ Ministral 8B (2410)Function callingMistral AI 100.0%
  2. 1≈ Llama 3.1 Nemotron Ultra 253B v1Function callingNVIDIA 100.0%
  3. 3 BitAgent Bounty 8BMode not statedBittensor 97.5%
  4. 4 Claude Haiku 4.5PromptingAnthropic 95.3%
  5. 5 Claude Sonnet 4.5PromptingAnthropic 95.0%
  6. 6 Gemini 2.5 FlashFunction callingGoogle 93.7%
  7. 7 Gemini 2.5 Flash-LitePromptingGoogle 93.3%
  8. 8 DeepSeek-V3.2-ExpFunction callingDeepSeek 93.2%
  9. 9 Gemini 2.5 Flash-LiteFunction callingGoogle 92.5%
  10. 10 Mistral Medium 3Function callingMistral AI 92.0%
  11. 11 Gemini 2.5 FlashPromptingGoogle 91.1%
  12. 12 GPT-5 miniFunction callingOpenAI 91.0%
  13. 13 ToolACE-2-8BFunction callingHuawei Noah's Ark Lab and USTC 90.8%
  14. 13 Claude Opus 4.5PromptingAnthropic 90.8%
  15. 15 Hammer2.1-7BFunction callingMadeAgents 90.1%
  16. 60 Grok 4.1 Fast (reasoning)Function callingxAI 79.4%

Full results

BFCL: irrelevance detection, % correct, higher is better
#ModelIrrelevance detection
% correct, higher is better
Price
$ per million tokens, in / out
1≈ Ministral 8B (2410)Function callingMistral AI
100.0%
–
1≈ Llama 3.1 Nemotron Ultra 253B v1Function callingNVIDIA
100.0%
–
3 BitAgent Bounty 8BMode not statedBittensor
97.5%
–
4 Claude Haiku 4.5PromptingAnthropic
95.3%
$1 / $5
5 Claude Sonnet 4.5PromptingAnthropic
95.0%
$3 / $15
6 Gemini 2.5 FlashFunction callingGoogle
93.7%
$0.30 / $2.50
7 Gemini 2.5 Flash-LitePromptingGoogle
93.3%
$0.10 / $0.40
8 DeepSeek-V3.2-ExpFunction callingDeepSeek
93.2%
$0.27 / $0.41
9 Gemini 2.5 Flash-LiteFunction callingGoogle
92.5%
$0.10 / $0.40
10 Mistral Medium 3Function callingMistral AI
92.0%
$0.40 / $2
11 Gemini 2.5 FlashPromptingGoogle
91.1%
$0.30 / $2.50
12 GPT-5 miniFunction callingOpenAI
91.0%
$0.25 / $2
13 ToolACE-2-8BFunction callingHuawei Noah's Ark Lab and USTC
90.8%
–
13 Claude Opus 4.5PromptingAnthropic
90.8%
$5 / $25
15 Hammer2.1-7BFunction callingMadeAgents
90.1%
–
16 GPT-5 nanoFunction callingOpenAI
89.1%
$0.05 / $0.40
17 Mistral Small 3.2 24BFunction callingMistral AI
87.9%
$0.094 / $0.25
18 Phi-4PromptingMicrosoft
87.6%
$0.07 / $0.14
19 Kimi K2 (0711)Function callingMoonshot AI
87.3%
$0.57 / $2.30
19 Falcon3 1B InstructFunction callingTII
87.3%
–
19 GPT-5.2PromptingOpenAI
87.3%
$1.75 / $14
22 Qwen3-14BPromptingAlibaba
87.2%
$0.10 / $0.22
22 o4-miniPromptingOpenAI
87.2%
–
24 CoALM 8BMode not statedUIUC and Oumi
86.9%
–
25 Command A ReasoningFunction callingCohere
86.8%
–
26 Claude Sonnet 4.5Function callingAnthropic
86.6%
$3 / $15
27 GPT-4.1Function callingOpenAI
86.5%
$2 / $8
28 o3Function callingOpenAI
86.1%
$2 / $8
28 Hammer2.1-3BFunction callingMadeAgents
86.1%
–
30 CoALM 70BMode not statedUIUC and Oumi
85.7%
–
31 Gemini 3 Pro PreviewPromptingGoogle
85.6%
–
32 Claude Haiku 4.5Function callingAnthropic
85.1%
$1 / $5
33 GLM-4.6 (reasoning on)Function callingZ.ai
85.0%
$0.50 / $2
34 Qwen3-4B-Instruct-2507Function callingAlibaba
84.9%
–
35 Claude Opus 4.5Function callingAnthropic
84.7%
$5 / $25
36 Grok 4 (0709)PromptingxAI
84.3%
–
37 Command AFunction callingCohere
84.2%
$2.50 / $10
38 GPT-4.1PromptingOpenAI
84.0%
$2 / $8
38 o3PromptingOpenAI
84.0%
$2 / $8
40 o4-miniFunction callingOpenAI
83.9%
–
41 GPT-4.1 nanoPromptingOpenAI
83.4%
$0.10 / $0.40
42 Nanbeige4-3B-Thinking (2511)Function callingNanbeige
83.1%
–
43 Qwen3-0.6BPromptingAlibaba
82.5%
–
44 RZN-TPromptingPhronetic AI
82.4%
–
44 Qwen3-32BPromptingAlibaba
82.4%
$0.14 / $0.40
46 Qwen3-8BPromptingAlibaba
82.3%
–
47 Arch-Agent-32BMode not statedKatanemo
82.2%
–
48 Nova 2 LiteFunction callingAmazon
82.1%
$0.30 / $2.50
49 Qwen3-14BFunction callingAlibaba
81.9%
$0.10 / $0.22
50 Qwen3-235B-A22B-Instruct-2507Function callingAlibaba
81.7%
$0.15 / $0.75
50 GPT-4.1 miniFunction callingOpenAI
81.7%
$0.40 / $1.60
50 Command R7B (12-2024)Function callingCohere
81.7%
$0.037 / $0.15
53 Palmyra X 004Function callingWriter
81.0%
–
54 Qwen3-0.6BFunction callingAlibaba
80.8%
–
54 Hammer2.1-0.5BFunction callingMadeAgents
80.8%
–
56 Granite 3.2 8B InstructFunction callingIBM
80.5%
–
57 xLAM-2-32b-fc-rFunction callingSalesforce
80.2%
–
58 Granite 3.1 8B InstructFunction callingIBM
80.0%
–
59 Qwen3-30B-A3B-Instruct-2507Function callingAlibaba
79.9%
$0.09 / $0.30
60 Grok 4.1 Fast (reasoning)Function callingxAI
79.4%
–
60 GPT-5.2Function callingOpenAI
79.4%
$1.75 / $14
60 Hammer2.1-1.5BFunction callingMadeAgents
79.4%
–
63 xLAM-2-70b-fc-rFunction callingSalesforce
79.1%
–
63 Qwen3-8BFunction callingAlibaba
79.1%
–
65 Qwen3-235B-A22B-Instruct-2507PromptingAlibaba
78.9%
$0.15 / $0.75
66 Gemini 3 Pro PreviewFunction callingGoogle
77.9%
–
67 Qwen3-1.7BFunction callingAlibaba
76.5%
–
68 Qwen3-32BFunction callingAlibaba
76.4%
$0.14 / $0.40
69 Qwen3-4B-Instruct-2507PromptingAlibaba
75.9%
–
70 Grok 4 (0709)Function callingxAI
75.4%
–
71 Granite 20B FunctionCallingFunction callingIBM
75.1%
–
72 Qwen3-30B-A3B-Instruct-2507PromptingAlibaba
74.9%
$0.09 / $0.30
73 Arch-Agent-1.5BMode not statedKatanemo
74.8%
–
74 Arch-Agent-3BMode not statedKatanemo
74.7%
–
75 Mistral Medium 3Mode not statedMistral AI
74.5%
$0.40 / $2
76 Nanbeige3.5-Pro-ThinkingFunction callingNanbeige
74.2%
–
77 Grok 4.1 Fast (non-reasoning)Function callingxAI
74.1%
–
78 GPT-4.1 miniPromptingOpenAI
73.9%
$0.40 / $1.60
79 MiniCPM3-4BPromptingOpenBMB
73.7%
–
79 Gemma 3 27BPromptingGoogle
73.7%
$0.08 / $0.35
81 MiniCPM3-4BFunction callingOpenBMB
72.8%
–
82 Nova Micro 1.0Function callingAmazon
70.7%
$0.035 / $0.14
83 Gemma 3 12BPromptingGoogle
70.3%
$0.05 / $0.15
84 Nova Pro 1.0Function callingAmazon
70.1%
$0.80 / $3.20
85 Mistral Large 2.1 (2411)Function callingMistral AI
68.9%
–
86 DeepSeek-V3.2-Exp (reasoning on)PromptingDeepSeek
67.0%
$0.27 / $0.41
87 GPT-4.1 nanoFunction callingOpenAI
66.0%
$0.10 / $0.40
88 Mistral Small 3.2 24BPromptingMistral AI
65.7%
$0.094 / $0.25
89 xLAM-2-1b-fc-rFunction callingSalesforce
64.5%
–
90 xLAM-2-3b-fc-rFunction callingSalesforce
63.5%
–
91 xLAM-2-8b-fc-rFunction callingSalesforce
63.3%
–
92 Mistral NemoFunction callingMistral AI
61.8%
$0.029 / $0.03
93 Granite 4.0 350MFunction callingIBM
60.8%
–
94 Llama 4 Maverick 17B 128E Instruct (FP8)Function callingMeta
56.0%
–
95 GPT-5 miniPromptingOpenAI
55.7%
$0.25 / $2
96 Gemma 3 4BPromptingGoogle
53.9%
$0.04 / $0.05
97 Llama 3.3 70B InstructFunction callingMeta
53.5%
$0.71 / $0.71
98 Llama 3.2 3B InstructFunction callingMeta
52.1%
$0.05 / $0.33
99 Llama 3.2 1B InstructFunction callingMeta
51.6%
$0.027 / $0.20
100 GPT-5 nanoPromptingOpenAI
45.8%
$0.05 / $0.40
101 Llama 4 ScoutFunction callingMeta
44.9%
$0.10 / $0.30
102 Llama 3.1 8B InstructPromptingMeta
42.7%
$0.05 / $0.08
103 Mistral Large 2.1 (2411)PromptingMistral AI
38.8%
–
104 Bielik 11B v2.3 InstructPromptingSpeakLeash and ACK Cyfronet AGH
36.0%
–
105 Gemma 3 1BPromptingGoogle
33.2%
–
106 Falcon3 3B InstructFunction callingTII
32.9%
–
107 Falcon3 10B InstructFunction callingTII
32.1%
–
108 Falcon3 7B InstructFunction callingTII
32.0%
–
109 Mistral NemoPromptingMistral AI
6.3%
$0.029 / $0.03

Swipe the table sideways for more columns.

Source: Berkeley Function Calling Leaderboard (BFCL) V4, via Spring Prompt, springprompt.com/benchmarks/bfcl. Test your own prompt on these models →

Ranks follow the score as shown, so equal numbers share a rank. Results as published by Berkeley Function Calling Leaderboard (BFCL) V4; we do not re-run them.

Progress over time

Irrelevance detection, plotted at each model's release date, higher is better. Each model counts once, at its best setting. Scores are today's, not launch day's; 38 models without a known release date are left off. Hover or tap a mark for its model.

Best score so far

Every model at its release date. The line moves each time a new release beats every earlier one, and each new best is named where there is room.

20%40%60%80%100%120% Jul 20242025Jul2026Jul Mistral NemoMinistral 8B (2410)

Best score by lab

Each lab's best so far, stepping whenever one of its releases beats its own earlier models. The 6 labs with the best scores today; 7 others not drawn.

Mistral AIAnthropicGoogleDeepSeekOpenAIMicrosoft

60%80%100% Jul 20242025Jul2026Jul Mistral AIAnthropicGoogleDeepSeekOpenAIMicrosoft

What it measures

Declining to call a function when none of those offered fits.

What it does not measure

Not general refusal or safety behaviour.

Berkeley Function Calling Leaderboard (BFCL) V4 by the UC Berkeley Gorilla team, Apache License 2.0.