GGUF models by ubergarm

42 GGUF models quantized by ubergarm on Hugging Face. Sizes run from 10.7B to 1030B parameters. The longest trained context is 384K tokens.

Most downloadedNewestLargestLongest contextSmallest
GI
GigaChat3 10B A1.8B Bf16
Gigachat · ubergarm
Params10.7B
Context256K
Smallest3.3 GB
GGUF
QW
Qwen3 30B A3B Instruct 2507
Qwen · ubergarm
Params30.5B
Context256K
Smallest8.1 GB
GGUF
QW
Qwen3.8 27B
Qwen · ubergarm
Params28.6B
Context256K
Smallest16.9 GB
GGUF
KI
Kimi K2 Thinking
Kimi · ubergarm
Params1030B
Context256K
Smallest235.1 GB
ReasoningTool use
QW
Qwen3 Coder 30B A3B Instruct
Qwen · ubergarm
Params30.5B
Context256K
Smallest8.1 GB
CodingTool use
QW
Qwen3.5 35B A3B
Qwen · ubergarm
Params34.7B
Context256K
Smallest21.2 GB
ReasoningTool use
QW
Qwen3 Coder Next
Qwen · ubergarm
Params79.7B
Context256K
Smallest20.5 GB
CodingTool use
QW
Qwen3 235B A22B Instruct 2507
Qwen · ubergarm
Params235B
Context256K
Smallest87.9 GB
ReasoningTool use
KI
Kimi K2 Instruct
Kimi · ubergarm
Params1030B
Context128K
Smallest1.5 GB
GGUF
DE
DeepSeek R1 0528
DeepSeek · ubergarm
Params672B
Context160K
Smallest139.8 GB
ReasoningTool use
GL
GLM 4.7
GLM · ubergarm
Params358B
Context198K
Smallest88.5 GB
ReasoningTool use
QW
Qwen3.6 27B
Qwen · ubergarm
Params26.9B
Context256K
Smallest15.8 GB
ReasoningTool use
QW
Qwen3.5 397B A17B
Qwen · ubergarm
Params396B
Context256K
Smallest1.1 GB
GGUF
KI
Kimi K2.6
Kimi · ubergarm
Params1030B
Context256K
Smallest235.1 GB
ReasoningTool use
QW
Qwen3.5 122B A10B
Qwen · ubergarm
Params122B
Context256K
Smallest32.5 GB
ReasoningTool use
GL
GLM 4.5 Air
GLM · ubergarm
Params110B
Context128K
Smallest38.7 GB
ReasoningTool use
MI
MiniMax M2.5
Minimax · ubergarm
Params229B
Context192K
Smallest75.0 GB
ReasoningTool use
GL
GLM 5
GLM · ubergarm
Params754B
Context198K
Smallest181.7 GB
ReasoningTool use
GL
GLM 4.7 Flash
GLM · ubergarm
Params29.9B
Context198K
Smallest16.0 GB
ReasoningTool use
GL
GLM 4.5
GLM · ubergarm
Params358B
Context128K
Smallest90.0 GB
ReasoningTool use
QW
Qwen3.5 27B
Qwen · ubergarm
Params26.9B
Context256K
Smallest16.6 GB
ReasoningTool use
MI
MiniMax M2.7
Minimax · ubergarm
Params229B
Context192K
Smallest47.2 GB
ReasoningTool use
GL
GLM 5.1
GLM · ubergarm
Params754B
Context198K
Smallest181.7 GB
ReasoningTool use
GL
GLM 4.6
GLM · ubergarm
Params357B
Context198K
Smallest86.9 GB
ReasoningTool use
ST
Step 3.5 Flash
Step · ubergarm
Params197B
Context256K
Smallest57.8 GB
ReasoningTool use
DE
DeepSeek TNG R1T2 Chimera
DeepSeek · ubergarm
Params672B
Context160K
Smallest142.7 GB
Tool use
DE
DeepSeek V3.2 Speciale Bf16 Safetensors
DeepSeek · ubergarm
Params671B
Context160K
Smallest156.9 GB
ReasoningTool use
KI
Kimi K2.5
Kimi · ubergarm
Params1030B
Context256K
Smallest235.1 GB
ReasoningTool use
LI
Ling 1T
Ling · ubergarm
Params1000B
Context32K
Smallest231.3 GB
ReasoningTool use
DE
DeepSeek V3 0324
DeepSeek · ubergarm
Params672B
Context160K
Smallest139.8 GB
Tool use

About ubergarm on Hugging Face

ubergarm has published 42 GGUF models that ggufmodels.com tracks, across families such as Qwen, GLM and Kimi. The most downloaded are GigaChat3 10B A1.8B Bf16, Qwen3 30B A3B Instruct 2507 and Qwen3.8 27B.

The quants ubergarm publishes most often are IQ1_S (4), IQ4_NL (3), Q8_0 (2) and Q4_0 (2). Parameter counts run from 10.7B to 1030B, and the longest trained context window is 384K tokens.

Frequently asked questions

What GGUF models has ubergarm published?

42 models are listed here, mainly from the Qwen, GLM and Kimi families. The most downloaded is GigaChat3 10B A1.8B Bf16.

Which quants does ubergarm publish?

The most common are IQ1_S (4 models), IQ4_NL (3 models), Q8_0 (2 models) and Q4_0 (2 models).

Keep browsing