YOUR MEMORY BUDGET

Local AI models for 8GB VRAM.

Explore models by theoretical weight storage. This shortlist covers available downloadable language models with a known or name-derived parameter count.

6 GBleft for weights at 4-bit
8 GB budget − 2 GB reserve

The reserve is your assumption, not a measured overhead allowance. The default 2 GB may be too little for your runtime or context length. Results are not confirmed fits and have no verified speed rating. Weight estimates use decimal GB and do not include quantization overhead. For timed results on one 12GB machine, see measured speeds.

852 models within the weight budget

Largest parameter count first · not a quality ranking

Tencent · Downloadable

WeDLM-8B-Base

Text generation

Generates text for conversations and writing.

Source listed
Dec 25, 2025
Parameters
8B
4-bit weights only
≈ 4 GB
License
apache-2.0
Compare this model →

Tongyi · Alibaba · Downloadable

MAI-UI-8B

Vision

Understands text and images and responds in text.

Source listed
Dec 25, 2025
Parameters
8B
4-bit weights only
≈ 4 GB
License
apache-2.0
Compare this model →

Z.ai · Downloadable

webrl-llama-3.1-8b

Text generation

Generates text, according to its published model architecture (LlamaForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.

Source listed
Nov 5, 2024
Parameters
8B
4-bit weights only
≈ 4 GB
License
other
Compare this model →

inclusionAI · Downloadable

Ling-3.0-tiny

Text generation

Generates text for conversations and writing.

Source listed
Aug 10, 2026
Parameters
7.9B
4-bit weights only
≈ 3.9 GB
License
mit
Compare this model →

Google · Downloadable

gemma-3n-E4B

VisionVideo understandingTranscription

Understands text and images and responds in text.

Source listed
Jun 3, 2025
Parameters
7.8B
4-bit weights only
≈ 3.9 GB
License
gemma
Compare this model →

Google · Downloadable

gemma-3n-E4B-it

VisionVideo understandingTranscription

Understands text and images and responds in text.

Source listed
Jun 3, 2025
Parameters
7.8B
4-bit weights only
≈ 3.9 GB
License
gemma
Compare this model →

OpenMOSS · Downloadable

Embodied_Planner-R1-Alfworld

Text generation

Generates text, according to its published model architecture (Qwen2ForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.

Source listed
Jun 27, 2025
Parameters
7.6B
4-bit weights only
≈ 3.8 GB
License
cc-by-4.0
Compare this model →

OpenMOSS · Downloadable

Embodied_R1-ScienceWorld

Text generation

Generates text, according to its published model architecture (Qwen2ForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.

Source listed
Jun 27, 2025
Parameters
7.6B
4-bit weights only
≈ 3.8 GB
License
cc-by-4.0
Compare this model →

BAAI · Downloadable

Matroyshka-ReRanker-beir

Text generation

Generates text, according to its published model architecture (CostWiseMistralForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.

Source listed
May 21, 2025
Parameters
7.1B
4-bit weights only
≈ 3.6 GB
License
Not disclosed
Compare this model →

BAAI · Downloadable

Matroyshka-ReRanker-document

Text generation

Generates text, according to its published model architecture (CostWiseMistralForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.

Source listed
May 21, 2025
Parameters
7.1B
4-bit weights only
≈ 3.6 GB
License
Not disclosed
Compare this model →

Aleph Alpha · Downloadable

llama-tfree-hat-pretrained-7b-dpo

Text generation

Generates text, according to its published model architecture (HATForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.

Source listed
Jul 30, 2025
Parameters
7B
4-bit weights only
≈ 3.5 GB
License
other
Compare this model →

What changes with a 8GB budget?

With 2 GB set aside, 6 GB remains for weights. At 4-bit that allows up to 12 billion parameters in the theoretical calculation. An 8-billion-parameter model alone needs approximately 4 GB at this precision, before its conversation cache and runtime.

Why might a listed model still fail to load?

Its actual weight format may need more space, the runtime may not support its architecture, or the conversation cache may exceed your reserve. Vision models can also need space for image processing. Check the original model card and measure your exact setup before relying on it.

Read the memory guide and technical sources · Compare two models