Multiple AST

BFCL · 50 models

Rank	Model	Score
1	Qwen3-32B (Prompt)	97
2	Qwen3-8B (FC)	96.5
3	Gemini-3-Pro-Preview (Prompt)	96
3	Mistral-Small-2506 (Prompt)	96
3	palmyra-x-004 (FC)	96
6	Claude-Opus-4-5-20251101 (FC)	95.5
6	Claude-Sonnet-4-5-20250929 (FC)	95.5
8	GLM-4.6 (FC thinking)	95
8	Llama-3.1-8B-Instruct (Prompt)	95
8	Llama-3.3-70B-Instruct (FC)	95
8	Llama-4-Maverick-17B-128E-Instruct-FP8 (FC)	95
8	Qwen3-235B-A22B-Instruct-2507 (Prompt)	95
8	Qwen3-8B (Prompt)	95
14	Arch-Agent-3B	94.5
14	Qwen3-30B-A3B-Instruct-2507 (FC)	94.5
14	Qwen3-32B (FC)	94.5
17	Amazon-Nova-2-Lite-v1:0 (FC)	94
17	Arch-Agent-32B	94
17	Claude-Haiku-4-5-20251001 (FC)	94
17	Gemini-3-Pro-Preview (FC)	94
17	Gemma-3-12b-it (Prompt)	94
17	Llama-4-Scout-17B-16E-Instruct (FC)	94
17	xLAM-2-32b-fc-r (FC)	94
17	xLAM-2-70b-fc-r (FC)	94
25	Claude-Opus-4-5-20251101 (Prompt)	93.5
25	CoALM-8B	93.5
25	Command A Reasoning (FC)	93.5
25	Falcon3-10B-Instruct (FC)	93.5
25	GPT-4.1-2025-04-14 (Prompt)	93.5
25	Grok-4-0709 (Prompt)	93.5
25	Hammer2.1-3b (FC)	93.5
25	Mistral-small-2506 (FC)	93.5
25	Nanbeige4-3B-Thinking-2511 (FC)	93.5
25	Qwen3-14B (Prompt)	93.5
25	Qwen3-4B-Instruct-2507 (FC)	93.5
25	mistral-large-2411 (FC)	93.5
25	xLAM-2-8b-fc-r (FC)	93.5
38	Amazon-Nova-Pro-v1:0 (FC)	93
38	BitAgent-Bounty-8B	93
38	Command A (FC)	93
38	Grok-4-1-fast-non-reasoning (FC)	93
38	Grok-4-1-fast-reasoning (FC)	93
38	Qwen3-14B (FC)	93
38	Qwen3-30B-A3B-Instruct-2507 (Prompt)	93
45	GPT-4.1-mini-2025-04-14 (Prompt)	92.5
45	Gemma-3-27b-it (Prompt)	92.5
45	Grok-4-0709 (FC)	92.5
45	Hammer2.1-7b (FC)	92.5
45	Llama-3.2-3B-Instruct (FC)	92.5
45	Open-Mistral-Nemo-2407 (FC)	92.5