IQ4_XS GGUF models

Importance-matrix 4-bit quant, extra small. Small files with good quality for the size. 97368 models have a IQ4_XS file. Sizes run from 1M to 2450B parameters. The longest trained context is 102400K tokens.

Most downloadedNewestLargestLongest contextSmallest
SA
Salamandra 7b Instruct
Salamandra · RichardErkhov
Params7.77B
Context8K
Q4_K_M4.9 GB
GGUF
CR
Crimson_Dawn v0.2 Base
Crimson · bartowski
Params12.2B
Context1000K
Q4_K_M7.5 GB
Roleplay
GE
Gemma 4 12B Coder Fable5 Composer2.5 v1 Uncensored Heretic
Gemma · mradermacher
Params11.9B
Context256K
Q4_K_M7.4 GB
CodingReasoningVision
QW
Qwen3-32B-128K
Qwen · unsloth
Params32.8B
Context128K
Q4_K_M19.8 GB
GGUF
QW
Qwen3 Embedding 0.6B
Qwen · PeterAM4
Params596M
Context32K
Q4_K_M0.4 GB
Embedding
QW
Qwen3.8 27B TURBO Fable Cold Fusion 735 882 Heretic Uncensored NEO CODER MAX MTP
Qwen · robertbschuette
Params27B
Context256K
Q4_K_M18.0 GB
CodingReasoningRoleplay
CO
ContextualKunoichi KTO 7B i1
Contextualkunoichi · mradermacher
Params7.24B
Context32K
Q4_K_M4.4 GB
GGUF
QW
Qwen2 0.5B NuminaMath TIR v1
Qwen · mradermacher
Params494M
Context32K
Q4_K_M0.4 GB
GGUF
LF
LFM2.5 8B A1B Uncensored
Lfm · zaakirio
Params8.47B
Context125K
Q4_K_M5.2 GB
GGUF
ST
stabilityai japanese stablelm base beta 70b
Stabilityai · RichardErkhov
Params1.6B
Context—
Q4_K_M0.0 GB
GGUF
QW
Qwen3.8 27B TURBO Fable Cold Fusion 735 882 Heretic Uncensored NEO CODER MAX MTP
Qwen · Code-Gen
Params26.9B
Context256K
Q4_K_M18.0 GB
CodingReasoningRoleplay
DE
DeepHat V1 7B imatrix
Deephat · liodon-ai
Params7.62B
Context32K
Q4_K_M4.7 GB
Coding
SK
Sky T1 32B Flash
Sky · bartowski
Params32.8B
Context32K
Q4_K_M19.9 GB
GGUF
ZE
zephyr 7b hh dpo i1
Zephyr · mradermacher
Params7.24B
Context32K
Q4_K_M4.4 GB
GGUF
LL
Llama 3.2 9B Uncensored Brainstorm Alpha
Llama · DavidAU
Params8.94B
Context128K
Q4_K_M5.4 GB
Roleplay
AP
Apriel 1.5 15b Thinker
Apriel · mradermacher
Params14.4B
Context256K
Q4_K_M8.8 GB
ReasoningVision
LL
lightblue suzume llama 3 8B japanese
Llama · mmnga
Params8B
Context8K
Q4_K_M4.9 GB
GGUF
QW
Qwen2.5 Wolverine CODER 11B V2 128k Ctx
Qwen · mradermacher
Params10.9B
Context128K
Q4_K_M6.6 GB
Coding
GR
Granite-4.0-Micro
Granite · unsloth
Params3.4B
Context128K
Q4_K_M2.1 GB
GGUF
NE
Nemotron 3.5 Bolt E4B Heretic
Nemotron · mradermacher
Params7.52B
Context64K
Q4_K_M5.3 GB
ReasoningTool use
DI
DiscoveryRSI 27B
Discoveryrsi · bartowski
Params27.3B
Context256K
Q4_K_M17.4 GB
ReasoningTool useVision
NI
Nitral AI Eris PrimeV4.20 Vision 32k 7B
Nitral · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
Vision
QW
Qwopus3.8 27B Flash V2 Abliterated
Qwopus · mradermacher
Params27B
Context256K
Q4_K_M16.5 GB
Vision
TE
TeaAI Nemo 12B
Teaai · mradermacher
Params12.2B
Context128K
Q4_K_M7.5 GB
Roleplay
QW
Qwen3 VL 2B Instruct heretic
Qwen · saidutta69
Params1.72B
Context256K
Q4_K_M1.1 GB
RoleplayTool useVision
QW
Qwen3.8 27B
Qwen · zugquants
Params26.9B
Context256K
Q4_K_M16.5 GB
CodingReasoningTool use
UM
Umt5 Xxl
Umt · chatpig
Params5.68B
Context512
Q4_K_M3.7 GB
Coding
SI
SIRI 7B High
Siri · mradermacher
Params7.62B
Context128K
Q4_K_M4.7 GB
GGUF
WA
wan s2v
Wan · calcuis
Params16.3B
Context—
Q4_K_M13.9 GB
GGUF
GE
Gemma 4 E4B It OBLITERATED
Gemma · mradermacher
Params7.46B
Context128K
Q4_K_M5.3 GB
GGUF

Other quants

Keep browsing