DeepSeek-R1
DeepSeek · China
DeepSeek reasoning model R1 · Open weights
- Release month
- Field
- Chat
- Availability
- open-weight
- Parameters
- 671B-A37B, Distill 1.5B·7B·8B·14B·32B·70B, R1-0528-Qwen3-8B
- License
- MIT · Commercial OK
- Size
- Distill 1.5B·7B·8B·14B·32B·70B · R1-0528-Qwen3-8B · 671B-A37B
Overview
- Listed on Hugging Face January 20, 2025 · Open weights
- Reasoning model · Includes 0528 update and small distilled versions · MIT
View official announcement ↗
← All models