GGUF models by lm-kit

109 GGUF models quantized by lm-kit on Hugging Face. Sizes run from 33M to 70.6B parameters. The longest trained context is 1024K tokens.

Most downloadedNewestLargestLongest contextSmallest
BG
bge m3
BGE · lm-kit
Params567M
Context8K
Q4_K_M0.4 GB
Embedding
QW
Qwen3 14B
Qwen · lm-kit
Params14.8B
Context40K
Q4_K_M9.0 GB
GGUF
GE
gemma 3 4b instruct
Gemma · lm-kit
Params3.88B
Context128K
Q4_K_M2.5 GB
GGUF
QW
Qwen3 8B
Qwen · lm-kit
Params8.19B
Context40K
Q4_K_M5.0 GB
GGUF
PH
phi 4 mini 3.8b instruct
Phi · lm-kit
Params3.84B
Context128K
Q4_K_M2.5 GB
GGUF
QW
Qwen3 0.6B
Qwen · lm-kit
Params752M
Context40K
Q4_K_M0.5 GB
ReasoningTool use
QW
Qwen3 1.7B
Qwen · lm-kit
Params2.03B
Context40K
Q4_K_M1.3 GB
GGUF
GE
gemma 3 1b instruct
Gemma · lm-kit
Params1B
Context32K
Q4_K_M0.8 GB
GGUF
PH
phi 3.5 mini 3.8b instruct
Phi · lm-kit
Params3.82B
Context128K
Q4_K_M2.4 GB
GGUF
HA
Harrier Oss v1 0.6b
Harrier · lm-kit
Params596M
Context32K
Q4_K_M0.4 GB
ReasoningTool use
EM
Embeddinggemma 300m
Embeddinggemma · lm-kit
Params303M
Context2K
Q4_K_M0.2 GB
Embedding
GE
gemma 3 270m instruct
Gemma · lm-kit
Params268M
Context32K
Q4_K_M0.3 GB
GGUF
GE
gemma 3 12b instruct
Gemma · lm-kit
Params11.8B
Context128K
Q4_K_M7.3 GB
GGUF
QW
Qwen3 4B
Qwen · lm-kit
Params4.02B
Context40K
Q4_K_M2.5 GB
ReasoningTool use
NO
nomic embed text 1.5
Nomic · lm-kit
Params137M
Context2K
Q4_K_M0.1 GB
Embedding
QW
qwen 2.5 7b instruct
Qwen · lm-kit
Params7.62B
Context32K
Q4_K_M4.7 GB
Tool use
GL
GLM 4.7 Flash
GLM · lm-kit
Params29.9B
Context198K
Q4_K_M18.1 GB
ReasoningTool use
QW
Qwen3 Embedding 0.6B
Qwen · lm-kit
Params596M
Context32K
Q4_K_M0.4 GB
EmbeddingReasoningTool use
PH
phi 3 medium 4k 14b instruct
Phi · lm-kit
Params14B
Context4K
Q4_K_M8.6 GB
GGUF
GE
gemma 2 27b
Gemma · lm-kit
Params27.2B
Context8K
Q4_K_M16.6 GB
GGUF
DO
docling-layout-heron
Docling · lm-kit
Params42M
Context—
Smallest0.2 GB
GGUF
QW
qwen 2.5 math 7.6b instruct
Qwen · lm-kit
Params7.62B
Context4K
Q4_K_M4.7 GB
Tool use
PH
phi 3.1 mini 4k 3.8b instruct
Phi · lm-kit
Params3.82B
Context4K
Q4_K_M2.4 GB
GGUF
GE
gemma 4 e2b instruct lmk
Gemma · lm-kit
Params4.65B
Context128K
Smallest9.3 GB
ReasoningTool useVision
QW
qwen 2.5 3.1b instruct
Qwen · lm-kit
Params3.09B
Context32K
Q4_K_M1.9 GB
Tool use
MI
mistral nemo 2407 12.2b instruct
Mistral · lm-kit
Params12.2B
Context1000K
Q4_K_M7.5 GB
Tool use
BG
bge m3 reranker v2
BGE · lm-kit
Params568M
Context8K
Q4_K_M0.4 GB
Embedding
LL
llama 3.2 3b instruct
Llama · lm-kit
Params3.21B
Context128K
Q4_K_M2.0 GB
Tool use
LL
llama 3.2 1b instruct
Llama · lm-kit
Params1.24B
Context128K
Q4_K_M0.8 GB
Tool use
PH
phi 4 14.7b instruct
Phi · lm-kit
Params14.7B
Context16K
Q4_K_M9.1 GB
GGUF

About lm-kit on Hugging Face

lm-kit has published 109 GGUF models that ggufmodels.com tracks, across families such as Qwen, Gemma and Mistral. The most downloaded are bge m3, Qwen3 14B and gemma 3 4b instruct.

The quants lm-kit publishes most often are F16 (90), Q4_K_M (66), Q8_0 (48) and Q6_K (48). Their Q4_K_M files range from 0.1 to 42.5 GB. Parameter counts run from 33M to 70.6B, and the longest trained context window is 1024K tokens.

Frequently asked questions

What GGUF models has lm-kit published?

109 models are listed here, mainly from the Qwen, Gemma and Mistral families. The most downloaded is bge m3.

Which quants does lm-kit publish?

The most common are F16 (90 models), Q4_K_M (66 models), Q8_0 (48 models) and Q6_K (48 models).

Keep browsing