DeepSeek-R1-Distill-Qwen-14B

14B

DeepSeek

R1 reasoning distilled into 14B. Huge community interest; excellent chain-of-thought.

⬇ 47.6K HF downloads♥ 236 likesbartowski/DeepSeek-R1-Distill-Qwen-14B-GGUF· stats from 6/24/2026

Consumer GPU

131K

Max Context

Quant Variants

EXL2 4.65bpw

Best Quality

98.0%

Accuracy Retained

Quantization Variants

Per-quant VRAM, quality loss, and inference speed on RTX 4090

Format	Level	BPW	VRAM	PPL Loss	Speed	Actions
GGUF	Q4_K_M	4.85	10.2 GB	2.8%	95 tok/s	Calc HF
EXL2	4.65bpw	4.65	9.8 GB	2.0%	128 tok/s	Calc HF
AWQ	INT4	4	9.2 GB	3.6%	118 tok/s	Calc HF