Tiny GGUF models
62726 GGUF models with 3B parameters or fewer. Sizes run from 1M to 3.5B parameters. The longest trained context is 795898K tokens.
QW
Dolphin3.0 Qwen2.5 3B
Qwen · Loverboy69
Params3.09B
Context32K
Q4_K_M1.9 GB
GR
Granite 3.0 2b Instruct
Granite · mradermacher
Params2.63B
Context4K
Q4_K_M1.6 GB
QW
QwenDGAF Coder 1.5B
Qwen · hecan3d
Params1.54B
Context32K
Q4_K_M1.0 GB
QW
Jev Style Qwen3.5 2B Decision v2
Qwen · chaoliangUNSW
Params1.88B
Context256K
Q4_K_M1.3 GB
GA
gaggle-nanochat-pretrained-climbmix-20260924
Gaggle · castorini
Params3.3B
Context2K
Q4_K_M2.6 GB
LF
LFM2.5 Encoder 350M Diffusion
Lfm · mradermacher
Params354M
Context125K
Q4_K_M0.2 GB
ST
stable cypher instruct 3b
Stable · ragraph-ai
Params2.8B
Context16K
Q4_K_M1.7 GB
SM
smollm1.7b slerp
SmolLM · mradermacher
Params1.71B
Context2K
Q4_K_M1.1 GB
QW
QwenDGAF Coder 0.5B
Qwen · hecan3d
Params494M
Context32K
Q4_K_M0.4 GB
GR
Granite-3.3-2B-Instruct
Granite · unsloth
Params2.53B
Context128K
Q4_K_M1.5 GB
MI
MiniCPM5 2B
MiniCPM · eaddario
Params2.52B
Context—
Smallest0.0 GB
AN
Anima turbo v1.0
Anima · Abiray
Params2.09B
Context—
Q4_K_M1.4 GB
OR
Orpheus 3b Chinese FT Q8 0
Orpheus · lex-au
Params3.3B
Context128K
Smallest3.5 GB
QW
Qwen3.5 2B Polaris HighIQ INSTRUCT
Qwen · mradermacher
Params1.88B
Context256K
Q4_K_M1.3 GB
QW
heretic_EVA Qwen2.5 1.5B v0.0
Qwen · mradermacher
Params1.54B
Context128K
Q4_K_M1.0 GB
QW
Qwen3.5 0.8B NVFP4 MTP
Qwen · michaelw9999
Params752M
Context256K
Smallest0.6 GB
DE
DeepSeek R1 Distill Qwen 1.5B Fully Uncensored
DeepSeek · 234-45Qq
Params1.78B
Context128K
Smallest1.3 GB
QW
dolphin-2.9.3-qwen2-0.5b
Qwen · mradermacher
Params494M
Context128K
Q4_K_M0.4 GB
VS
VST 3B
Vst · mradermacher
Params3.4B
Context125K
Q4_K_M2.1 GB
QW
SatellaRAG 0.8B
Qwen · EphAsad
Params773M
Context256K
Q4_K_M0.5 GB
CO
codec.cpp
Codec · BricksDisplay
Params74M
Context—
Q4_K_M0.1 GB
BI
BiRefNet_lite
Birefnet · Acly
Params220M
Context—
Smallest0.1 GB
IR
Irodori-TTS-v4-Large
Irodori · sakasegawa
Params3.38B
Context—
Smallest2.6 GB
LL
Layerskip Llama3.2 1B
Llama · matrixportalx
Params1.24B
Context128K
Q4_K_M0.8 GB
LL
Llama 3.2 1b Chatml RP i1
Llama · mradermacher
Params1.5B
Context128K
Q4_K_M1.0 GB
RU
RusMind Malysh 4.7
Rusmind · DICPlay
Params752M
Context256K
Q4_K_M0.5 GB
WA
Lynx
Wan · vantagewithai
Params2.1B
Context—
Q4_K_M0.2 GB
EU
euclaise ReMask 3B
Euclaise · RichardErkhov
Params2.8B
Context4K
Q4_K_M1.7 GB
MI
MiniChat-1.5-3B
Minichat · mradermacher
Params3.02B
Context4K
Q4_K_M1.8 GB
QW
Qwen2.5 Math 1.5B Instruct
Qwen · legraphista
Params1.54B
Context4K
Smallest0.4 GB