Berkeley Function Calling Leaderboard leaderboard: who is #1

Name: Berkeley Function Calling Leaderboard rankings
Creator: UC Berkeley Gorilla
License: https://creativecommons.org/licenses/by/4.0/

UC Berkeley Gorilla · Overall accuracy, higher is better · 81 models · published Apr 12, 2026

Top 25 of 81 models, best variant per model as published

#	Model	Score
1	Claude Opus 4.5as “Claude-Opus-4-5-20251101 (FC)”	77.5%
2	Claude Sonnet 4.5as “Claude-Sonnet-4-5-20250929 (FC)”	73.2%
3	Gemini 3 Proas “Gemini-3-Pro-Preview (Prompt)”	72.5%
4	GLM 4.6as “GLM-4.6 (FC thinking)”	72.4%
5	Grok 4 1 Fastas “Grok-4-1-fast-reasoning (FC)”	69.6%
6	Claude 4.5 Haikuas “Claude-Haiku-4-5-20251001 (FC)”	68.7%
7	o3as “o3-2025-04-16 (Prompt)”	63.0%
8	Grok 4as “Grok-4-0709 (Prompt)”	63.0%
9	Moonshotai Kimi K2as “Moonshotai-Kimi-K2-Instruct (FC)”	59.1%
10	Grok 4 1 Fast Nonas “Grok-4-1-fast-non-reasoning (FC)”	58.3%
11	Command A+as “Command A Reasoning (FC)”	57.1%
12	DeepSeek-V3.2as “DeepSeek-V3.2-Exp (Prompt + Thinking)”	56.7%
13	Gemini 2.5 Flash (Sep 2025)as “Gemini-2.5-Flash (FC)”	56.2%
14	GPT-5.2as “GPT-5.2-2025-12-11 (FC)”	55.9%
15	GPT-5 minias “GPT-5-mini-2025-08-07 (FC)”	55.5%
16	Xlam 2 32b Fc Ras “xLAM-2-32b-fc-r (FC)”	54.7%
17	GPT 4.1as “GPT-4.1-2025-04-14 (FC)”	54.0%
18	o4-minias “o4-mini-2025-04-16 (FC)”	53.2%
19	Xlam 2 70b Fc Ras “xLAM-2-70b-fc-r (FC)”	53.1%
20	Qwen3-235B-A22B-Thinking (Jul 2025)as “Qwen3-235B-A22B-Instruct-2507 (Prompt)”	52.1%
21	GPT 5 nanoas “GPT-5-nano-2025-08-07 (FC)”	51.5%
22	Nanbeige4 3bas “Nanbeige4-3B-Thinking-2511 (FC)”	51.4%
23	GPT 4.1 minias “GPT-4.1-mini-2025-04-14 (FC)”	50.5%
24	Qwen3 32bas “Qwen3-32B (FC)”	48.7%
25	Nanbeige3.5 Proas “Nanbeige3.5-Pro-Thinking (FC)”	47.7%