GGUF models by DevQuasar-12

44 GGUF models quantized by DevQuasar-12 on Hugging Face. Sizes run from 1B to 671B parameters. The longest trained context is 10240K tokens.

Most downloadedNewestLargestLongest contextSmallest
QW
Featherless ai.Qwerky QwQ 32B
Qwen · DevQuasar-12
Params34.9B
Context1024K
Q4_K_M22.8 GB
Reasoning
LL
Meta llama.Llama 4 Scout 17B 16E Instruct
Llama · DevQuasar-12
Params108B
Context10240K
Q4_K_M65.4 GB
Tool use
LL
nvidia/Llama-3.1-8B-UltraLong-1M-Instruct
Llama · DevQuasar-12
Params8.04B
Context1048K
Q4_K_M4.9 GB
Tool use
DE
DeepScaleR 1.5B Preview
Deepscaler · DevQuasar-12
Params1.78B
Context128K
Q4_K_M1.1 GB
Tool use
GL
Huihui ai.GLM 4 9B 0414 Abliterated
GLM · DevQuasar-12
Params9.4B
Context32K
Q4_K_M6.2 GB
Tool use
LL
Huihui ai.Llama 3.1 Nemotron Nano 8B v1 Abliterated
Llama · DevQuasar-12
Params8.03B
Context128K
Q4_K_M4.9 GB
GGUF
GE
Google.Gemma 3 4b It Qat Int4 Unquantized
Gemma · DevQuasar-12
Params3.88B
Context128K
Q4_K_M2.5 GB
Vision
GL
Huihui ai.GLM Z1 9B 0414 Abliterated
GLM · DevQuasar-12
Params9.4B
Context32K
Q4_K_M6.2 GB
Tool use
GR
ibm-granite/granite-3.3-2b-instruct
Granite · DevQuasar-12
Params2.53B
Context128K
Q4_K_M1.5 GB
ReasoningTool use
GL
THUDM/GLM-4-9B-0414
GLM · DevQuasar-12
Params9.4B
Context32K
Q4_K_M6.2 GB
Tool use
GR
ibm-granite/granite-3.3-2b-base
Granite · DevQuasar-12
Params2.53B
Context128K
Q4_K_M1.5 GB
GGUF
MI
NousResearch/Nous-Hermes-2-Mistral-7B-DPO
Mistral · DevQuasar-12
Params7.24B
Context32K
Q4_K_M4.4 GB
Tool use
LL
meta-llama/Llama-4-Maverick-17B-128E-Instruct
Llama · DevQuasar-12
Params401B
Context1024K
Q4_K_M242.8 GB
Tool use
LI
inclusionAI.Ling Coder Lite
Ling · DevQuasar-12
Params16.8B
Context4K
Q4_K_M11.2 GB
Coding
GE
Google.Gemma 3 12b It Qat Int4 Unquantized
Gemma · DevQuasar-12
Params11.8B
Context128K
Q4_K_M7.3 GB
Vision
OP
All Hands.Openhands Lm 32b v0.1
Openhands · DevQuasar-12
Params32.8B
Context32K
Q4_K_M19.9 GB
Tool use
GE
Google.Gemma 3 1b It Qat Int4 Unquantized
Gemma · DevQuasar-12
Params1B
Context32K
Q4_K_M0.8 GB
GGUF
QW
Huihui ai.Qwen2.5 32B Instruct Abliterated SFT
Qwen · DevQuasar-12
Params32.8B
Context32K
Q4_K_M19.9 GB
Tool use
LL
Llama 3.2 3B
Llama · DevQuasar-12
Params3.61B
Context128K
Q4_K_M2.2 GB
ReasoningTool use
LI
Saxo.Linkbricks Horizon AI Korean Superb 22B
Linkbricks · DevQuasar-12
Params22.2B
Context32K
Q4_K_M13.3 GB
Tool use
GR
ibm-granite/granite-3.3-8b-instruct
Granite · DevQuasar-12
Params8.17B
Context128K
Q4_K_M4.9 GB
ReasoningTool use
QW
Deepcogito.Cogito v1 Preview Qwen 14B
Qwen · DevQuasar-12
Params14.8B
Context128K
Q4_K_M9.0 GB
ReasoningTool use
HA
Delta Vector.Hamanasu Magnum 4B
Hamanasu · DevQuasar-12
Params4.51B
Context128K
Q4_K_M2.8 GB
GGUF
GL
THUDM/GLM-4-32B-Base-0414
GLM · DevQuasar-12
Params32.6B
Context32K
Q4_K_M19.7 GB
GGUF
QW
VIDraft.QwQ R1984 32B
Qwen · DevQuasar-12
Params32.8B
Context128K
Q4_K_M19.9 GB
ReasoningTool use
CI
Sao10K.70B L3.3 Cirrus X1
Cirrus · DevQuasar-12
Params70.6B
Context128K
Q4_K_M42.5 GB
Tool use
OP
All Hands.Openhands Lm 32b v0.1 Ep3
Openhands · DevQuasar-12
Params32.8B
Context32K
Q4_K_M19.9 GB
Tool use
HA
Delta Vector.Hamanasu 4B Instruct KTO V2
Hamanasu · DevQuasar-12
Params4.51B
Context128K
Q4_K_M2.8 GB
GGUF
LL
Llama 3 8B Instruct Coding Expert
Llama · DevQuasar-12
Params8.03B
Context8K
Q4_K_M4.9 GB
GGUF
DE
Deepseek ai.DeepSeek V3 0324 Bf16 2
DeepSeek · DevQuasar-12
Params671B
Context160K
Q4_K_M404.4 GB
Tool use

About DevQuasar-12 on Hugging Face

DevQuasar-12 has published 44 GGUF models that ggufmodels.com tracks, across families such as Llama, Qwen and GLM. The most downloaded are Featherless ai.Qwerky QwQ 32B, Meta llama.Llama 4 Scout 17B 16E Instruct and nvidia/Llama-3.1-8B-UltraLong-1M-Instruct.

The quants DevQuasar-12 publishes most often are Q4_K_M (44), Q2_K (44), Q5_K_M (43) and Q8_0 (42). Their Q4_K_M files range from 0.8 to 404.4 GB. Parameter counts run from 1B to 671B, and the longest trained context window is 10240K tokens.

Frequently asked questions

What GGUF models has DevQuasar-12 published?

44 models are listed here, mainly from the Llama, Qwen and GLM families. The most downloaded is Featherless ai.Qwerky QwQ 32B.

Which quants does DevQuasar-12 publish?

The most common are Q4_K_M (44 models), Q2_K (44 models), Q5_K_M (43 models) and Q8_0 (42 models).

Keep browsing