Q3_K GGUF models
A GGUF quantization format. Lower numbers mean smaller files and more quality loss. 15171 models have a Q3_K file. Sizes run from 1M to 5632B parameters. The longest trained context is 102400K tokens.
SP
Speechless Hermes Coig Lite 13b
Speechless · RichardErkhov
Params13B
Context4K
Q4_K_M7.9 GB
LL
zomd AISquare Instruct llama2 koen 13b v0.9.17
Llama · RichardErkhov
Params13.2B
Context2K
Q4_K_M8.0 GB
XU
XueyingJia pythia 1b deduped hh online dpo merge
Xueyingjia · RichardErkhov
Params1.01B
Context2K
Q4_K_M0.7 GB
QW
Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive
Qwen · odexus
Params34.7B
Context256K
Q4_K_M21.2 GB
TI
TinySolar-248m-4k-code-instruct
Tinysolar · RichardErkhov
Params248M
Context32K
Q4_K_M0.2 GB
QW
Qwen2-0.5B-OnlineDPO-PairRM
Qwen · RichardErkhov
Params494M
Context32K
Q4_K_M0.4 GB
QW
Qwen2.5 72B Instruct
Qwen · RichardErkhov
Params72.7B
Context32K
Q4_K_M47.4 GB
MI
Mistral-7b-instruct-v0.2-summ-sft-e2m
Mistral · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
MI
mlfoundations dev mistral alpaca sft sample 5
Mistral · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
MI
sania963 Mistral instruct sql v0.3
Mistral · RichardErkhov
Params7.25B
Context32K
Q4_K_M4.4 GB
WE
Weyaxi openchat 3.5 1210 Seraph Slerp
Weyaxi · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
IN
InternLM2
InternLM · RichardErkhov
Params19.9B
Context8K
Q4_K_M12.0 GB
MI
InnerI InnerILLM OpenPipe Nous Yarn Mistral optimized 1228 7B slerp
Mistral · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
BR
BroAlanTaps GPT2 large 128 46000steps
Broalantaps · RichardErkhov
Params838M
Context1K
Q4_K_M0.5 GB
MI
Moriacrafter Mistral 7B v0.2 none DepressionDetection Reddit
Mistral · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
JI
JiuZhang3.0 Synthesis 7B
Jiuzhang · RichardErkhov
Params6.91B
Context4K
Q4_K_M4.2 GB
MI
EleutherAI Mistral 7B v0.1 population first ft
Mistral · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
MI
Yarn-Mistral-7b-128k-DPO
Mistral · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
LL
Meta Llama 3 8B Instruct
Llama · RichardErkhov
Params8.03B
Context8K
Q4_K_M4.9 GB
QW
Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive
Qwen · Qweassdzxc
Params34.7B
Context256K
Q4_K_M21.2 GB
GE
Google Gemma 2 9b It
Gemma · RichardErkhov
Params9.24B
Context8K
Q4_K_M5.8 GB
FA
gguf falcon mamba 7b instruct
Falcon · Felladrin
Params7.27B
Context1024K
Smallest2.8 GB
MI
OpenHermes 2.5 Mistral 7B
Mistral · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
RE
reflect_llm8B_Om2SftT1-Om2IpsdpG8kIpsdpIter1T1_b0.1
Reflect · RichardErkhov
Params8.03B
Context128K
Q4_K_M4.9 GB
LL
Llama 3.2 Adr Full Model
Llama · RichardErkhov
Params1.24B
Context128K
Q4_K_M0.8 GB
LL
mlfoundations dev llama2 oh teknium scaling down random 0.8
Llama · RichardErkhov
Params6.74B
Context4K
Q4_K_M4.1 GB
ME
mergekit community mergekit slerp janrfan
Mergekit · RichardErkhov
Params8B
Context128K
Q4_K_M4.9 GB
LL
Llama 3 ELYZA JP 8B
Llama · RichardErkhov
Params8.03B
Context8K
Q4_K_M4.9 GB
LL
mlfoundations dev llama3 1 8b oh v3.1 wo unnatural instructions
Llama · RichardErkhov
Params8.03B
Context128K
Q4_K_M4.9 GB
AY
aya 101
Aya · kcoopermiller
Params12.9B
Context—
Smallest4.2 GB