GGUF models that fit in 16 GB

182247 GGUF language models have at least one quant whose estimated memory use is 16 GB or less at 8K context. Each card shows the largest quant that fits. Sizes run from 1M to 20260731B parameters. The longest trained context is 4194304K tokens.

Most downloadedNewestLargestLongest contextSmallest
QW
Qwen3.8-27B
Qwen · unsloth
Params27.3B
Context256K
Best fitQ3_K_XL 15.8 GB
Vision
OR
Ornith-1.5-9B
Ornith · ornith-ai
Params9.2B
Context256K
Best fitQ8_0 11.4 GB
Vision
QW
Qwen3.8-27B
Qwen · huihui-ai
Params27.3B
Context256K
Best fitQ3_K_XL 16.0 GB
Vision
OR
Tiel Coder 35B A3B GGUF MTP
Ornith · peculiar-ragdoll
Params447M
Context256K
Best fitIQ3_S 15.3 GB
CodingVision
QW
Qwen3.8 27B Uncensored
Qwen · JonathanColetti
Params27.3B
Context256K
Best fitIQ2_M 13.3 GB
Vision
OR
Ornith 1.0 9B
Ornith · ornith-ai
Params8.95B
Context256K
Best fitQ8_0 11.1 GB
GGUF
MI
MiniMax H3
Minimax · unsloth
Params20B
Context—
Best fitQ2_K_M 15.2 GB
GGUF
QW
Qwen3.8-27B
Qwen · ggml-org
Params3B
Context256K
Best fitBF16 8.6 GB
Vision
MX
mxbai-embed-large-v1
Mxbai · mixedbread-ai
Params334M
Context512
Best fitF16 2.0 GB
Embedding
QW
Ternary Bonsai 2 27B
Qwen · prism-ml
Params27B
Context256K
Smallest5.9 GB
Vision
QW
Qwen3.8 27B
Qwen · byteshape
Params27.3B
Context256K
Best fitIQ4_XS 15.8 GB
Vision
QW
Qwen3.5 9B The Defiant Fable Uncensored Heretic NEO IMATRIX MAX MTP
Qwen · DavidAU
Params8.95B
Context256K
Best fitQ8_0 12.3 GB
CodingReasoningRoleplay
QW
Qwen3.8 27B TURBO Fable Cold Fusion 735 882 Heretic Uncensored NEO CODER MAX MTP
Qwen · DavidAU
Params27B
Context256K
Best fitIQ2_M 14.8 GB
CodingReasoningRoleplay
OR
Ornith 1.5 9B Uncensored
Ornith · mradermacher
Params8.95B
Context256K
Best fitQ8_0 11.1 GB
Vision
QW
Qwen Image 2.1
Qwen · unsloth
Params7.12B
Context—
Best fitQ8_0 9.1 GB
GGUF
GE
Gemma-4-12B-It
Gemma · unsloth
Params11.9B
Context256K
Best fitBF16 14.2 GB
Vision
QW
Huihui Qwen3.6 35B A3B Abliterated
Qwen · huihui-ai
Params35.5B
Context256K
Best fitQ2_K 14.4 GB
Vision
QW
Swift 1.5 Qwen3.8 27B GSQ RCO
Qwen · ukisai
Params27B
Context—
Best fitIQ3_S 14.1 GB
Reasoning
QW
Qwen3.5-9B
Qwen · unsloth
Params8.95B
Context256K
Best fitQ8_K_XL 14.5 GB
Vision
LA
Laguna XS 2.1 APEX
Laguna · mudler
Params33.4B
Context256K
Smallest12.8 GB
Coding
QW
Qwen3.8-27B
Qwen · peculiar-ragdoll
Params27.3B
Context256K
Best fitQ3_K_XL 15.8 GB
Vision
QW
Qwen3.6-35B-A3B
Qwen · unsloth
Params34.7B
Context256K
Best fitIQ3_S 14.8 GB
Vision
QW
Qwen3.6-35B-A3B
Qwen · unsloth
Params35.5B
Context256K
Best fitIQ3_XXS 15.3 GB
Vision
QW
Qwen3.8 27B GSQ RCO
Qwen · ISTA-DASLab
Params27B
Context—
Best fitIQ3_S 14.1 GB
Vision
QW
Qwen3.5-4B
Qwen · unsloth
Params4.21B
Context256K
Best fitBF16 10.0 GB
Vision
QW
Qwen3.8 35B A3B Distill
Qwen · empero-ai
Params35.5B
Context256K
Best fitQ2_K 15.0 GB
ReasoningVision
QW
Qwen3Vl 8b Instruct
Qwen · Qwen
Params8.19B
Context256K
Best fitQ8_0 10.4 GB
Vision
WA
Wan2.1
Wan · DeepBeepMeep
Params9.2B
Context256K
Best fitIQ3_S 15.3 GB
GGUF
GE
Gemma-4-26B-A4B-It
Gemma · unsloth
Params25.2B
Context256K
Best fitBF16 9.4 GB
Vision
HY
HyperCLOVAX SEED Text Instruct 1.5B
Hyperclovax · rippertnt
Params1.81B
Context128K
Best fitQ4_K_M 2.4 GB
GGUF

What runs in 16 GB of VRAM?

A GGUF model fits when its file plus the KV cache for an 8K context window plus about 0.5 GB of runtime stays under 16 GB. 182,247 language models have at least one quant that does. Each card shows the largest quant of that model that still fits, because the largest one that fits is the closest to the original quality.

The most downloaded models that fit are Qwen3.8-27B, Ornith-1.5-9B and Qwen3.8-27B. At this size, models up to roughly 14B fit at 4-bit, and 7B and 8B models can use 6-bit or 8-bit quants. If you are close to the limit, a smaller context or a smaller quant frees memory.

Frequently asked questions

What is the best GGUF model for 16 GB of VRAM?

By downloads, Qwen3.8-27B is the most popular model that fits. “Best” depends on what you do, so compare the sizes and quants on this page and check the model pages for the details.

How is "fits in 16 GB" decided?

We add the GGUF file size, an fp16 KV cache for an 8K context (worked out from each model’s header) and 0.5 GB for the runtime. The result is an estimate, not a measurement.

Other memory sizes

Keep browsing