GGUF models by AmpereComputing

68 GGUF models quantized by AmpereComputing on Hugging Face. Sizes run from 494M to 671B parameters. The longest trained context is 10240K tokens.

Most downloadedNewestLargestLongest contextSmallest
QW
Qwen3Vl 4b Instruct
Qwen · AmpereComputing
Params4.02B
Context256K
Smallest4.1 GB
Tool useVision
DE
Deepseek-R1-0528
DeepSeek · AmpereComputing
Params671B
Context160K
Smallest441.6 GB
ReasoningTool use
QW
Qwen3Vl 8b Instruct
Qwen · AmpereComputing
Params8.19B
Context256K
Smallest8.3 GB
Tool useVision
GE
Gemma 3 4b It
Gemma · AmpereComputing
Params3.88B
Context128K
Smallest3.9 GB
Vision
QW
Qwen2.5-Vl-7B-Instruct
Qwen · AmpereComputing
Params7.62B
Context125K
Smallest7.7 GB
Vision
QW
Qwen2.5 VL 3B Instruct
Qwen · AmpereComputing
Params3.09B
Context125K
Smallest3.1 GB
Vision
LL
Meta Llama 3.1 8B
Llama · AmpereComputing
Params8.03B
Context128K
Smallest5.0 GB
GGUF
GE
Gemma 3 27b It
Gemma · AmpereComputing
Params27B
Context128K
Smallest27.7 GB
Vision
QW
qwen 3 a3b 30b
Qwen · AmpereComputing
Params30.5B
Context40K
Smallest18.9 GB
ReasoningTool use
GE
Gemma-3-12B-It
Gemma · AmpereComputing
Params11.8B
Context128K
Smallest12.1 GB
Vision
LL
llama 4 scout 16e 17b instruct
Llama · AmpereComputing
Params108B
Context10240K
Smallest108.6 GB
Tool use
LL
Llama 3.2 1B Instruct
Llama · AmpereComputing
Params1.24B
Context128K
Smallest0.8 GB
Tool use
QW
Qwen3Vl 2b Instruct
Qwen · AmpereComputing
Params1.72B
Context256K
Smallest1.7 GB
Tool useVision
QW
qwen 3 coder a35b 480b instruct
Qwen · AmpereComputing
Params480B
Context256K
Smallest484.1 GB
CodingTool use
AF
AFM 4.5B
Afm · AmpereComputing
Params4.62B
Context64K
Smallest4.7 GB
GGUF
MI
mistral v0.3 7b instruct
Mistral · AmpereComputing
Params7.25B
Context32K
Smallest4.5 GB
GGUF
QW
Qwen3 32B
Qwen · AmpereComputing
Params32.8B
Context40K
Smallest33.0 GB
ReasoningTool use
QW
qwen 3 a22b 235b
Qwen · AmpereComputing
Params235B
Context128K
Smallest145.2 GB
ReasoningTool use
TR
Trinity Nano Preview
Trinity · AmpereComputing
Params6.12B
Context128K
Smallest6.2 GB
Tool use
BG
BAAI/bge-m3
BGE · AmpereComputing
Params567M
Context8K
Smallest0.5 GB
Embedding
GL
Glm-4.5-Air
GLM · AmpereComputing
Params110B
Context128K
Smallest112.2 GB
ReasoningTool use
LL
Llama 3.2 3B Instruct
Llama · AmpereComputing
Params3.21B
Context128K
Smallest2.1 GB
Tool use
ER
ernie 4.5 a3b 21b thinking
Ernie · AmpereComputing
Params21.8B
Context128K
Smallest22.0 GB
ReasoningTool use
QW
Qwen2.5 14B Instruct
Qwen · AmpereComputing
Params14.8B
Context32K
Smallest14.9 GB
Tool use
PH
phi 4
Phi · AmpereComputing
Params14.7B
Context16K
Smallest14.8 GB
GGUF
GP
Gpt-Oss-20B
gpt-oss · AmpereComputing
Params20.9B
Context128K
Smallest22.2 GB
Tool use
BI
bielik v3.0 1.5b instruct
Bielik · AmpereComputing
Params1.6B
Context8K
Smallest1.6 GB
GGUF
GR
Granite 4.0 H Tiny
Granite · AmpereComputing
Params6.94B
Context1024K
Smallest4.7 GB
Tool use
DE
DeepSeek R1 Distill Qwen 14B
DeepSeek · AmpereComputing
Params14.8B
Context128K
Smallest14.9 GB
ReasoningTool use
QW
Qwen2.5 1.5B Instruct
Qwen · AmpereComputing
Params1.54B
Context32K
Smallest1.6 GB
Tool use

About AmpereComputing on Hugging Face

AmpereComputing has published 68 GGUF models that ggufmodels.com tracks, across families such as Qwen, DeepSeek and Llama. The most downloaded are Qwen3Vl 4b Instruct, Deepseek-R1-0528 and Qwen3Vl 8b Instruct.

The quants AmpereComputing publishes most often are Q4_K (9). Parameter counts run from 494M to 671B, and the longest trained context window is 10240K tokens.

Frequently asked questions

What GGUF models has AmpereComputing published?

68 models are listed here, mainly from the Qwen, DeepSeek and Llama families. The most downloaded is Qwen3Vl 4b Instruct.

Which quants does AmpereComputing publish?

The most common are Q4_K (9 models).

Keep browsing