utDeepSeek · Downloadable
deepseek-llm-7b-chat
Generates text for conversations and writing.
- Source listed
- Nov 29, 2023
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- other
YOUR MEMORY BUDGET
Local AI models for 12GB VRAM.
Explore models by theoretical weight storage. This shortlist covers available downloadable language models with a known or name-derived parameter count.
10 GBleft for weights at 4-bit
12 GB budget − 2 GB reserve
12 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.
978 models within the weight budget
Largest parameter count first · not a quality rankingDeepSeek · Downloadable
deepseek-math-7b-base
Generates text for conversations and writing.
- Source listed
- Feb 5, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- other
DeepSeek · Downloadable
deepseek-math-7b-instruct
Generates text for conversations and writing.
- Source listed
- Feb 5, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- other
DeepSeek · Downloadable
deepseek-math-7b-rl
Generates text for conversations and writing.
- Source listed
- Feb 5, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- other
DeepSeek · Downloadable
DeepSeek-Prover-V2-7B
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
- Apr 30, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- Not disclosed
DeepSeek · Downloadable
DeepSeek-R1-Distill-Qwen-7B
Generates text for conversations and writing. Supports a reasoning mode.
- Source listed
- Jan 20, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- mit
DeepSeek · Downloadable
deepseek-vl-7b-chat
Understands text and images and responds in text.
- Source listed
- Mar 7, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- other
Google · Downloadable
codegemma-7b
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Mar 21, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- gemma
Google · Downloadable
codegemma-7b-it
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Mar 21, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- gemma
Google · Downloadable
gemma-1.1-7b-it
Generates text for conversations and writing.
- Source listed
- Mar 26, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- gemma
Google · Downloadable
gemma-7b
Generates text for conversations and writing.
- Source listed
- Feb 8, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- gemma
Google · Downloadable
gemma-7b-it
Generates text for conversations and writing.
- Source listed
- Feb 13, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- gemma
inclusionAI · Downloadable
AReaL-boba-RL-7B
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
- Mar 29, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
inclusionAI · Downloadable
ASearcher-Local-7B
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
- Aug 6, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
inclusionAI · Downloadable
ASearcher-Web-7B
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
- Aug 5, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
inclusionAI · Downloadable
LLaDA-MoE-7B-A1B-Base
Generates text for conversations and writing.
- Source listed
- Sep 10, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
inclusionAI · Downloadable
UI-Venus-Ground-7B
Understands text and images and responds in text.
- Source listed
- Aug 15, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
inclusionAI · Downloadable
UI-Venus-Navi-7B
Understands text and images and responds in text.
- Source listed
- Aug 16, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
inclusionAI · Downloadable
ZwZ-7B
Understands text and images and responds in text.
- Source listed
- Feb 12, 2026
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
JetBrains · Downloadable
CodeLlama-7B-Kexer
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- May 13, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
JetBrains · Downloadable
CodeLlama-7B-KStack
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- May 13, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
JetBrains · Downloadable
CodeLlama-7B-KStack-clean
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- May 13, 2024
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- apache-2.0
M-A-P · Downloadable
CriticLeanGPT-Qwen2.5-7B-Instruct-SFT
Generates text for conversations and writing.
- Source listed
- Jul 11, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- other
M-A-P · Downloadable
CriticLeanGPT-Qwen2.5-7B-Instruct-SFT-RL
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
- Jul 11, 2025
- Parameters
- 7B
- 4-bit weights only
- ≈ 3.5 GB
- License
- Not disclosed
What changes with a 12GB budget?
With 2 GB set aside, 10 GB remains for weights. At 4-bit that allows up to 20 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