GGUF models by DevQuasar-6

110 GGUF models quantized by DevQuasar-6 on Hugging Face. Sizes run from 362M to 102B parameters. The longest trained context is 1000K tokens.

Most downloadedNewestLargestLongest contextSmallest
QW
Qwen.Qwen2.5 VL 3B Instruct
Qwen · DevQuasar-6
Params669M
Context125K
Q4_K_M1.9 GB
Vision
ET
EtherealAurora 12B
Etherealaurora · DevQuasar-6
Params12.2B
Context1000K
Q4_K_M7.5 GB
GGUF
QW
Qwen.Qwen2.5 VL 7B Instruct
Qwen · DevQuasar-6
Params677M
Context125K
Q4_K_M4.7 GB
Vision
QW
Qwen.Qwen2.5 VL 32B Instruct
Qwen · DevQuasar-6
Params689M
Context125K
Q4_K_M19.9 GB
Vision
MI
Mrfakename.Mistral Small 3.1 24b Instruct 2503 Hf
Mistral · DevQuasar-6
Params23.6B
Context32K
Q4_K_M14.3 GB
GGUF
LL
Trendmicro ailab.Llama Primus Reasoning
Llama · DevQuasar-6
Params8.03B
Context128K
Q4_K_M4.9 GB
Reasoning
RE
lightblue.lb reranker 0.5B v1.0
Reranker · DevQuasar-6
Params494M
Context32K
Q4_K_M0.4 GB
EmbeddingTool use
OP
All Hands.Openhands Lm 7b v0.1
Openhands · DevQuasar-6
Params7.62B
Context32K
Q4_K_M4.7 GB
Tool use
OP
knoveleng.OpenRS GRPO
Openrs · DevQuasar-6
Params1.78B
Context128K
Q4_K_M1.1 GB
ReasoningTool use
QW
Sao10K.14B Qwen2.5 Kunou v1
Qwen · DevQuasar-6
Params14.8B
Context128K
Q4_K_M9.0 GB
Tool use
GR
Ibm Granite.Granite 3.1 2b Instruct
Granite · DevQuasar-6
Params2.53B
Context128K
Q4_K_M1.5 GB
Tool use
CO
CohereForAI.c4ai Command R7B 12 2024
Command R · DevQuasar-6
Params8.03B
Context8K
Q4_K_M5.1 GB
Tool use
HU
FreedomIntelligence.HuatuoGPT O1 8B
Huatuogpt · DevQuasar-6
Params8.03B
Context128K
Q4_K_M4.9 GB
Tool use
SM
prithivMLmods.SmolLM2 CoT 360M
SmolLM · DevQuasar-6
Params362M
Context8K
Q4_K_M0.3 GB
GGUF
LL
NousResearch.Hermes 3 Llama 3.2 3B
Llama · DevQuasar-6
Params3.21B
Context128K
Q4_K_M2.0 GB
Tool use
LL
prithivMLmods.Llama Thinker 3B Preview
Llama · DevQuasar-6
Params3.21B
Context128K
Q4_K_M2.0 GB
ReasoningTool use
FA
tiiuae.Falcon3 10B Instruct
Falcon · DevQuasar-6
Params10.3B
Context32K
Q4_K_M6.3 GB
GGUF
FA
tiiuae.Falcon3 7B Instruct
Falcon · DevQuasar-6
Params7.46B
Context32K
Q4_K_M4.6 GB
GGUF
DE
Deci.DeciLM 7B
Decilm · DevQuasar-6
Params7.04B
Context8K
Q4_K_M4.2 GB
GGUF
LL
Llama 3.2 3B Instruct
Llama · DevQuasar-6
Params3.21B
Context128K
Q4_K_M2.0 GB
Tool use
TR
prithivMLmods.Triangulum 5B It
Triangulum · DevQuasar-6
Params5.41B
Context8K
Q4_K_M3.4 GB
GGUF
GR
Ibm Granite.Granite 3.1 1b A400M Base
Granite · DevQuasar-6
Params1.33B
Context128K
Q4_K_M0.8 GB
GGUF
LL
ruliad.deepthought 8b llama v0.01 alpha
Llama · DevQuasar-6
Params8.03B
Context128K
Q4_K_M4.9 GB
Tool use
GR
Ibm Granite.Granite 3.1 3b A800M Instruct
Granite · DevQuasar-6
Params3.3B
Context128K
Q4_K_M2.0 GB
Tool use
LL
Llama 3.2 1B Instruct
Llama · DevQuasar-6
Params1.24B
Context128K
Q4_K_M0.8 GB
Tool use
FI
HuggingFaceTB.finemath Ablation Finemath 3plus
Finemath · DevQuasar-6
Params3.21B
Context4K
Q4_K_M2.0 GB
GGUF
DE
Deci.DeciLM 6b Instruct
Decilm · DevQuasar-6
Params5.72B
Context4K
Q4_K_M3.4 GB
GGUF
MI
mistralai.Mistral Small 3.1 24B Base 2503
Mistral · DevQuasar-6
Params23.6B
Context128K
Q4_K_M14.3 GB
GGUF
MI
mistralai.Mistral Small 3.1 24B Instruct 2503
Mistral · DevQuasar-6
Params23.6B
Context128K
Q4_K_M14.3 GB
GGUF
TE
openGPT X.Teuken 7B Instruct Research v0.4
Teuken · DevQuasar-6
Params7.45B
Context4K
Q4_K_M5.0 GB
GGUF

About DevQuasar-6 on Hugging Face

DevQuasar-6 has published 110 GGUF models that ggufmodels.com tracks, across families such as Llama, Finemath and Qwen. The most downloaded are Qwen.Qwen2.5 VL 3B Instruct, EtherealAurora 12B and Qwen.Qwen2.5 VL 7B Instruct.

The quants DevQuasar-6 publishes most often are Q8_0 (110), Q5_K_M (110), Q4_K_M (110) and Q3_K_M (110). Their Q4_K_M files range from 0.3 to 61.3 GB. Parameter counts run from 362M to 102B, and the longest trained context window is 1000K tokens.

Frequently asked questions

What GGUF models has DevQuasar-6 published?

110 models are listed here, mainly from the Llama, Finemath and Qwen families. The most downloaded is Qwen.Qwen2.5 VL 3B Instruct.

Which quants does DevQuasar-6 publish?

The most common are Q8_0 (110 models), Q5_K_M (110 models), Q4_K_M (110 models) and Q3_K_M (110 models).

Keep browsing