GGUF models by DevQuasar-5

95 GGUF models quantized by DevQuasar-5 on Hugging Face. Sizes run from 135M to 479B parameters. The longest trained context is 1000K tokens.

Most downloadedNewestLargestLongest contextSmallest
GR
huihui ai.granite vision 3.2 2b abliterated
Granite · DevQuasar-5
Params2.53B
Context16K
Q4_K_M1.5 GB
Tool useVision
GR
Huihui Ai.Granite 3.2 2b Instruct Abliterated
Granite · DevQuasar-5
Params2.53B
Context128K
Q4_K_M1.5 GB
Tool use
LL
Llama 3.3 70B Instruct Abliterated
Llama · DevQuasar-5
Params70.6B
Context128K
Q4_K_M42.5 GB
Tool use
GR
Huihui Ai.Granite 3.2 8b Instruct Abliterated
Granite · DevQuasar-5
Params8.17B
Context128K
Q4_K_M4.9 GB
Tool use
GR
Granite 3b Code Instruct 128k
Granite · DevQuasar-5
Params3.48B
Context125K
Q4_K_M2.1 GB
Coding
GR
ibm granite.granite vision 3.1 2b preview
Granite · DevQuasar-5
Params2.53B
Context16K
Q4_K_M1.5 GB
Tool useVision
QW
Huihui ai.Qwen2.5 72B Instruct Abliterated
Qwen · DevQuasar-5
Params72.7B
Context32K
Q4_K_M47.4 GB
Tool use
OL
open r1.OlympicCoder 7B
Olympiccoder · DevQuasar-5
Params7.62B
Context32K
Q4_K_M4.7 GB
CodingReasoningTool use
GR
Huihui Ai.Granite 3.1 2b Instruct Abliterated
Granite · DevQuasar-5
Params2.53B
Context128K
Q4_K_M1.5 GB
Tool use
GE
Google.Gemma 3 4b Pt
Gemma · DevQuasar-5
Params3.88B
Context128K
Q4_K_M2.5 GB
Vision
CA
calme-3.2-instruct-78b
Calme · DevQuasar-5
Params78B
Context32K
Q4_K_M50.7 GB
GGUF
LL
nvidia.OpenMath2 Llama3.1 8B
Llama · DevQuasar-5
Params8.03B
Context128K
Q4_K_M4.9 GB
GGUF
SM
HuggingFaceTB.SmolLM2 135M Instruct
SmolLM · DevQuasar-5
Params135M
Context8K
Q4_K_M0.1 GB
GGUF
DE
Huihui ai.DeepSeekR1 QwQ SkyT1 32B Fusion 715
DeepSeek · DevQuasar-5
Params32.8B
Context128K
Q4_K_M19.9 GB
ReasoningTool use
EU
Utter project.EuroLLM 9B Instruct
Eurollm · DevQuasar-5
Params9.15B
Context4K
Q4_K_M5.6 GB
GGUF
GE
Google.Gemma 3 1b Pt
Gemma · DevQuasar-5
Params1B
Context32K
Q4_K_M0.8 GB
GGUF
GR
Ibm Granite.Granite 3.0 3b A800M Instruct
Granite · DevQuasar-5
Params3.37B
Context4K
Q4_K_M2.1 GB
Tool use
GR
Huihui Ai.Granite 3.1 8b Instruct Abliterated
Granite · DevQuasar-5
Params8.17B
Context128K
Q4_K_M4.9 GB
Tool use
BL
Bigscience.Bloomz 7b1 P3
Bloomz · DevQuasar-5
Params8.1B
Context2K
Q4_K_M5.3 GB
GGUF
XL
xLAM 7b R
Xlam · DevQuasar-5
Params7.24B
Context32K
Q4_K_M4.4 GB
Tool use
HP
Hpcgroup.Hpc Coder v2 6.7b
Hpc · DevQuasar-5
Params6.74B
Context16K
Q4_K_M4.1 GB
Coding
GR
ibm granite.granite vision 3.2 2b
Granite · DevQuasar-5
Params2.53B
Context16K
Q4_K_M1.5 GB
Tool useVision
LL
Orenguteng.Llama 3.1 8B Lexi Uncensored V2
Llama · DevQuasar-5
Params8.03B
Context128K
Q4_K_M4.9 GB
Tool use
BI
ContactDoctor.Bio Medical 3B CoT 012025
Bio · DevQuasar-5
Params3.09B
Context32K
Q4_K_M1.9 GB
GGUF
LL
neuralmagic.Sparse Llama 3.1 8B Ultrachat_200K 2of4
Llama · DevQuasar-5
Params8.03B
Context128K
Q4_K_M4.9 GB
Tool use
GR
Ibm Granite.Granite 3.1 8b Instruct
Granite · DevQuasar-5
Params8.17B
Context128K
Q4_K_M4.9 GB
Tool use
HU
Huihui ai.EXAONE 3.5 7.8B Instruct Abliterated
Huihui · DevQuasar-5
Params7.82B
Context32K
Q4_K_M4.8 GB
GGUF
NE
Nemotron Mini 4B Instruct
Nemotron · DevQuasar-5
Params4.19B
Context4K
Q4_K_M2.7 GB
Tool use
MO
allura org.MoE Girl 400MA 1BT
Moe · DevQuasar-5
Params1.33B
Context8K
Q4_K_M0.8 GB
GGUF
LL
ContactDoctor.Bio Medical Llama 3 8B
Llama · DevQuasar-5
Params8.03B
Context8K
Q4_K_M4.9 GB
GGUF

About DevQuasar-5 on Hugging Face

DevQuasar-5 has published 95 GGUF models that ggufmodels.com tracks, across families such as Llama, Granite and Qwen. The most downloaded are huihui ai.granite vision 3.2 2b abliterated, Huihui Ai.Granite 3.2 2b Instruct Abliterated and Llama 3.3 70B Instruct Abliterated.

The quants DevQuasar-5 publishes most often are Q8_0 (94), Q6_K (94), Q5_K_M (94) and Q4_K_M (94). Their Q4_K_M files range from 0.1 to 142.5 GB. Parameter counts run from 135M to 479B, and the longest trained context window is 1000K tokens.

Frequently asked questions

What GGUF models has DevQuasar-5 published?

95 models are listed here, mainly from the Llama, Granite and Qwen families. The most downloaded is huihui ai.granite vision 3.2 2b abliterated.

Which quants does DevQuasar-5 publish?

The most common are Q8_0 (94 models), Q6_K (94 models), Q5_K_M (94 models) and Q4_K_M (94 models).

Keep browsing