← Model library
text
DeepSeek R1 Distill 14B
DeepSeek · Family DeepSeek
The 14B distilled variant of DeepSeek R1 delivers near-large-model results on math, logic and planning with visible chain-of-thought. Runs comfortably on a single RTX 4090.
Key Features
- ✓ Visible chain-of-thought reasoning
- ✓ MIT license
- ✓ Runs on a single RTX 4090
Use Cases
- → Math and logic problem solving
- → Code review and explanation
- → Planning and decision support