GGUF models by DevQuasar

833 GGUF models quantized by DevQuasar on Hugging Face. Sizes run from 82M to 2780B parameters. The longest trained context is 10240K tokens.

Most downloadedNewestLargestLongest contextSmallest
QW
Qwen.Qwen3 VL Embedding 2B
Qwen · DevQuasar
Params1.72B
Context256K
Q4_K_M1.1 GB
EmbeddingVision
QW
Huihui ai.Qwen3 8B Abliterated
Qwen · DevQuasar
Params8.19B
Context40K
Q4_K_M5.0 GB
GGUF
DE
Deepseek ai.DeepSeek V3.2 BF16
DeepSeek · DevQuasar
Params670B
Context160K
Q4_K_M404.5 GB
GGUF
UN
baidu.Unlimited OCR
Unlimited · DevQuasar
Params2.93B
Context—
Q4_K_M2.0 GB
GGUF
SA
Sarvamai.Sarvam 30b
Sarvam · DevQuasar
Params32.2B
Context128K
Smallest12.2 GB
GGUF
LF
LiquidAI.LFM2 2.6B Transcript
Lfm · DevQuasar
Params2.57B
Context125K
Q4_K_M1.6 GB
GGUF
IN
amd.Instella MoE 16B A3B Think
Instella · DevQuasar
Params15.9B
Context32K
Smallest6.5 GB
Reasoning
GP
Openai.Gpt Oss 120b
gpt-oss · DevQuasar
Params117B
Context128K
Q4_K_M87.9 GB
Tool use
DE
Deepseek ai.DeepSeek V3.1 Terminus BF16
DeepSeek · DevQuasar
Params671B
Context160K
Q4_K_M404.5 GB
ReasoningTool use
GL
Zai org.GLM 5.3 Flash
GLM · DevQuasar
Params313B
Context1024K
Q4_K_M189.0 GB
Vision
KA
Kakaocorp.Kanana 1.5 8b Instruct 2505
Kanana · DevQuasar
Params8.03B
Context32K
Q4_K_M4.9 GB
Tool use
PH
microsoft.Phi 4 Reasoning Vision 15B
Phi · DevQuasar
Params14.7B
Context32K
Smallest5.5 GB
ReasoningVision
KA
Kakaocorp.Kanana 1.5 2.1b Instruct 2505
Kanana · DevQuasar
Params2.32B
Context32K
Q4_K_M1.5 GB
Tool use
MI
dphn.Dolphin Mistral 24B Venice Edition
Mistral · DevQuasar
Params23.6B
Context32K
Q4_K_M14.3 GB
GGUF
GL
Zai org.GLM 5
GLM · DevQuasar
Params754B
Context198K
Q4_K_M454.9 GB
ReasoningTool use
QW
Qwen.Qwen3.8 Flash Next
Qwen · DevQuasar
Params177B
Context256K
Smallest80.4 GB
ReasoningTool useVision
QW
Huihui ai.Qwen3 30B A3B Abliterated
Qwen · DevQuasar
Params30.5B
Context40K
Q4_K_M18.6 GB
GGUF
ER
baidu.ERNIE 4.5 300B A47B PT
Ernie · DevQuasar
Params300B
Context128K
Q4_K_M180.2 GB
GGUF
GE
Google.Gemma 4 E2B It Qat Q4_0 Unquantized
Gemma · DevQuasar
Params4.63B
Context128K
Smallest3.0 GB
ReasoningTool use
GL
zai org.GLM 5.2
GLM · DevQuasar
Params753B
Context1024K
Q4_K_M454.6 GB
ReasoningTool use
QW
Qwen.Qwen3 VL 235B A22B Instruct
Qwen · DevQuasar
Params235B
Context256K
Q4_K_M142.2 GB
Tool useVision
GR
Xai Org.Grok 2
Grok · DevQuasar
Params270B
Context128K
Q4_K_M164.1 GB
GGUF
MU
Meta models.Muse Glimmer 30B Assistant
Muse · DevQuasar
Params27.9B
Context128K
Q4_K_M1.5 GB
ReasoningTool useVision
QW
nicoboss.Qwen3 32B Uncensored
Qwen · DevQuasar
Params32.8B
Context40K
Q4_K_M19.8 GB
ReasoningTool use
DE
Deepseek ai.DeepSeek V3.1 F16
DeepSeek · DevQuasar
Params671B
Context160K
Q4_K_M404.5 GB
ReasoningTool use
CO
CohereLabs.command A Translate 08 2025
Command · DevQuasar
Params111B
Context8K
Q4_K_M67.1 GB
Tool use
QW
Huihui ai.Huihui Qwen3.5 27B Abliterated
Qwen · DevQuasar
Params26.9B
Context256K
Q4_K_M16.5 GB
ReasoningTool use
GP
Openai.Gpt Oss 20b
gpt-oss · DevQuasar
Params20.9B
Context128K
Q4_K_M15.8 GB
Tool use
KI
moonshotai.Kimi K2 Instruct 0905
Kimi · DevQuasar
Params1030B
Context256K
Q4_K_M620.6 GB
Tool use
QW
Qwen.Qwen3 Coder 30B A3B Instruct
Qwen · DevQuasar
Params30.5B
Context256K
Q4_K_M18.6 GB
CodingTool use

About DevQuasar on Hugging Face

DevQuasar has published 833 GGUF models that ggufmodels.com tracks, across families such as Qwen, Llama and Nemotron. The most downloaded are Qwen.Qwen3 VL Embedding 2B, Huihui ai.Qwen3 8B Abliterated and Deepseek ai.DeepSeek V3.2 BF16.

The quants DevQuasar publishes most often are Q4_K_M (762), Q2_K (758), Q3_K_M (755) and Q5_K_M (753). Their Q4_K_M files range from 0.1 to 1481.4 GB. Parameter counts run from 82M to 2780B, and the longest trained context window is 10240K tokens.

Frequently asked questions

What GGUF models has DevQuasar published?

833 models are listed here, mainly from the Qwen, Llama and Nemotron families. The most downloaded is Qwen.Qwen3 VL Embedding 2B.

Which quants does DevQuasar publish?

The most common are Q4_K_M (762 models), Q2_K (758 models), Q3_K_M (755 models) and Q5_K_M (753 models).

Keep browsing