IQ3_XXS GGUF models

A GGUF quantization format. Lower numbers mean smaller files and more quality loss. 28129 models have a IQ3_XXS file. Sizes run from 2M to 2450B parameters. The longest trained context is 10240K tokens.

Most downloadedNewestLargestLongest contextSmallest
RO
RoGemma-7b-Instruct
Rogemma · legraphista
Params8.54B
Context8K
Smallest2.2 GB
GGUF
QW
Qwen2-7B
Qwen · legraphista
Params7.62B
Context128K
Smallest1.9 GB
GGUF
LL
Llama 3.3 8B Thinking Gemini Flash 11000x 128k
Llama · mradermacher
Params8B
Context128K
Q4_K_M4.9 GB
ReasoningRoleplay
GE
Big Tiger Gemma 27B v3
Gemma · bartowski
Params28.4B
Context128K
Q4_K_M17.3 GB
Vision
CO
Command A Plus 05 2026
Command · bartowski
Params218B
Context195K
Q4_K_M135.1 GB
GGUF
MI
MixtralOrochi8x7B i1
Mistral · mradermacher
Params46.7B
Context32K
Q4_K_M28.6 GB
GGUF
IN
Internlm2 Math Base 20b
InternLM · mradermacher
Params19.9B
Context8K
Q4_K_M12.0 GB
GGUF
MA
Magnum v4 22b
Magnum · mradermacher
Params22.2B
Context32K
Q4_K_M13.3 GB
GGUF
KA
Kataguru Industry MVP 27B v1.1
Kataguru · kataguru
Params26.9B
Context256K
Q4_K_M2.0 GB
ReasoningTool useVision
GR
Granite 4.2 30b
Granite · bartowski
Params29.3B
Context128K
Q4_K_M18.0 GB
ReasoningTool use
DE
DeepSeek-Coder-V2-Lite-Instruct
DeepSeek · legraphista
Params15.7B
Context160K
Smallest5.0 GB
Coding
GL
GLM 5.3 Flash
GLM · batiai
Params313B
Context1024K
Q4_K_M188.7 GB
GGUF
MI
Nous-Hermes-2-Mixtral-8x7B-SFT
Mistral · mradermacher
Params46.7B
Context32K
Q4_K_M28.4 GB
Tool use
EX
EXAONE 3.5 2.4B Instruct 2
Exaone · ThomasBaruzier
Params2.67B
Context32K
Q4_K_M1.6 GB
GGUF
JA
JapanRailwayLLM
Japanrailwayllm · mradermacher
Params9.2B
Context8K
Q4_K_M5.8 GB
GGUF
LL
CodeLlama 7b Hf
Llama · mradermacher
Params6.74B
Context16K
Q4_K_M4.1 GB
Coding
GE
Gemma 4 E4B IT ABLITERATED UNCENSORED PHILADELPHIA CLASS
Gemma · mradermacher
Params7.46B
Context128K
Q4_K_M5.3 GB
GGUF
QW
Qwen2.5 Coder 14B Instruct 3MPER0RR Abliterated
Qwen · mradermacher
Params14.8B
Context32K
Q4_K_M9.0 GB
Coding
BO
Boptruth Agatha 7B i1
Boptruth · mradermacher
Params7.24B
Context128K
Q4_K_M4.4 GB
GGUF
MI
Midnight Miqu 70B v1.0 i1
Midnight · mradermacher
Params69B
Context32K
Q4_K_M41.7 GB
GGUF
EM
Emotion Companion 7b
Emotion · mradermacher
Params7.62B
Context32K
Q4_K_M4.7 GB
GGUF
LL
Fireball Meta Llama 3.2 8B Instruct agent 003 128k code DPO i1
Llama · mradermacher
Params8B
Context128K
Q4_K_M4.9 GB
Coding
QW
Qwen2.5 7B Instruct Heretic
Qwen · mradermacher
Params7.62B
Context32K
Q4_K_M4.7 GB
GGUF
PH
Phi 4 Reasoning Plus
Phi · bartowski
Params14.7B
Context32K
Q4_K_M9.1 GB
CodingReasoning
QW
Qwen3.6 35B A3B Abliterated
Qwen · mradermacher
Params34.7B
Context256K
Q4_K_M21.2 GB
GGUF
QW
Qwen3.5 4B RpRMax v1
Qwen · mradermacher
Params4.21B
Context256K
Q4_K_M2.7 GB
GGUF
GE
Gemma 4 31B It Heretic
Gemma · NeuralDreamer
Params30.7B
Context256K
Q4_K_M18.7 GB
CodingVision
CY
Cydonia 24B v4
Cydonia · bartowski
Params23.6B
Context128K
Q4_K_M14.3 GB
GGUF
IC
IceEspressoRPv6 i1
Iceespressorpv · mradermacher
Params7.24B
Context32K
Q4_K_M4.4 GB
GGUF
GP
gpt2 i1
Gpt · mradermacher
Params163M
Context1K
Q4_K_M0.1 GB
GGUF

Other quants

Keep browsing