HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks

Home/Orgs/Qwen/Qwen/Qwen-14B

Qwen/Qwen-14B

family qwen · 3 variants · unknown

provider: Qwen
released: —
huggingface: https://huggingface.co/Qwen/Qwen-14B
best scores: MMLU: 63.6C-Eval: 69.0GSM8K: 60.0MATH: 3.3HumanEval: 48.2MBPP: 20.8BBH: 38.2CMMLU: 31.8

Best scores per benchmark

benchmark	category	best score
C-Eval	knowledge	69.0
MMLU	knowledge	63.6
GSM8K	math	60.0
HumanEval	coding	48.2
BBH	reasoning	38.2
CMMLU	knowledge	31.8
MBPP	coding	20.8
MATH	math	3.3

Variants (3)

variant	kind	quant	bits	vRAM	dl	org	year
Qwen-14B	original	bf16	16.0		2,043	Qwen	2023
Qwen-14B-Chat-Int4	gptq	gptq			330	Qwen	2023
Qwen-14B-Chat-Int8	gptq	gptq			72	Qwen	2023

Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.