DeepSeek Models — Hardware Requirements
5 DeepSeek models from DeepSeek and the community, from the smallest that runs in 1.8 GB of VRAM up to 684.5B parameters. Every row links to full quantization tables and GPU compatibility.
All DeepSeek Models by Size
| Model | Params | Runs from | Context | Publisher | Quant downloads |
|---|---|---|---|---|---|
| DeepSeek OCR | 3.3B | 1.8 GB | 8K | ||
| Deepseek Vl2 Tiny | 3.4B | 7.4 GB | — | ||
| DeepSeek OCR 2 | 3.4B | 1.9 GB | 8K | ||
| Deepseek Llm 7B Base | 7B | 4.3 GB | 4K | ||
| Deepseek Llm 7B Chat | 7B | 4.3 GB | 4K | ||
| Deepseek Moe 16B Base | 16.4B | 7.7 GB | 4K | ||
| Deepseek Moe 16B Chat | 16.4B | 7.7 GB | 4K | ||
| DeepSeek TNG R1T2 Chimera | 684.5B | 192.1 GB | 164K |
Frequently Asked Questions
- How much VRAM do I need to run a DeepSeek model?
- The smallest DeepSeek model, DeepSeek OCR, runs from 1.8 GB of VRAM at an aggressive quantization. Larger family members need proportionally more — see the table above for every model.
- Which DeepSeek models can I run on a 16 GB GPU?
- 7 of 8 DeepSeek models fit in 16 GB of VRAM at some quantization, including DeepSeek OCR 2, DeepSeek OCR, Deepseek Vl2 Tiny.
- What is the most popular DeepSeek model to run locally?
- DeepSeek OCR 2 is the most downloaded DeepSeek model in local-friendly quantized formats. It runs from 1.9 GB of VRAM.