Function Calling on Berkeley Function Call Leaderboard (BFCL) Non-Live Out-of-Domain benchmark leaderboard · SOTA2 Research

Function Calling on Berkeley Function Call Leaderboard (BFCL) Non-Live Out-of-Domain

81.4AST SimpleHammer2.1-3B

Evaluation Results

Method	Links
Hammer2.1-3B2026.02		81.4	95	89.5	81.5	82.9	92	84	77.5	79
Hammer2.1-7B2026.02		78.1	95	93.5	88	86.4	92	86	77.5	81.6
GPT-4o2026.02		77.2	93.5	93	86	88.3	92	94	82.5	82.6
Qwen3-8B2026.02		76.8	95.5	94.5	88.5	94	92	86	77.5	83.4
ToolACE-8B2026.02		76.7	93.5	90.5	89.5	97.4	94	88	77.5	81
BigAgent-8B2026.02		76.2	95	94	82.5	98.6	94	88	77.5	81.6
Qwen3-4B2026.02		75.3	96.5	92	90.5	88	90	70	75	83.3
HyFunc♣2026.02		75.3	93.5	89.5	89	95	92	82	75	80.1
GPT-4o-mini2026.02		74.8	92	90	84	83.3	92	84	75	79.9
Hammer2.1-1.5B2026.02		74.7	92	84.5	80	86.6	90	82	75	75.5
xLAM-7B-r2026.02		74.2	95.5	81	73.5	74	96	82	67.5	76.6
xLAM-8x7B-r2026.02		73.6	90	69	38	89	90	72	45	74.3
Qwen2.5-3B-Instruct2026.02		73.3	92	73.5	76.5	86.9	90	66	70	74.9
Qwen2.5-32B-Instruct2026.02		72.8	94	93.5	88.5	97.6	88	84	77.5	82.2
Granite-20B2026.02		72.8	91.5	84	81.5	84.9	92	86	82.5	78.6
HyFunc♥2026.02		72.5	92.5	87.5	85	91	90	78	75	78.6
Qwen2.5-1.5B-Instruct2026.02		72.4	87	81.5	75.5	88	90	78	72.5	72.7
Qwen2.5-7B-Instruct2026.02		71.8	95	90	86	95.4	94	84	77.5	79.6
xLAM-1B-r2026.02		71.7	86	5	2	77.8	90	4	0	51.2
Qwen3-1.7B2026.02		71.1	93	87.5	81.5	87	88	68	75	76.9
GPT-4-turbo2026.02		70.4	91	90	87.5	87.4	90	86	77.5	81
Qwen2.5-14B-Instruct2026.02		69.7	95	88	89	90.4	92	72	85	79
Hammer2.1-0.5B2026.02		68	83	71.5	54	68.4	84	82	47.5	63.5
GLM4-9B-chat2026.02		65.2	81.5	0	0	94	90	0	0	55.1
Qwen3-0.6B2026.02		62.3	88	69	68	54	88	56	60	62.2
Qwen2.5-0.5B-Instruct2026.02		61.2	78	60	50	51.2	88	52	52.5	52.4