Model · DeepSeek · Reasoning
DeepSeek R1
DeepSeek's open-weight reasoning model. Released with full weights and a permissive MIT license — the first competitive open reasoning model.
- Modality
- Reasoning
- License
- MIT (Open)
- Parameter size
- 671B (37B active)
- Context window
- 128,000 tokens
- Released
- January 20, 2025
- Last verified
- May 10, 2026
- Runs locally
- Yes
Strengths
- MIT-licensed weights — fully self-hostable
- Competitive with closed reasoning models on math and coding
- Distilled smaller variants (1.5B–70B) for local hardware
Weaknesses
- Full model needs massive deployment footprint
- Reasoning chains can be verbose
Try it
| Where | Type | Notes |
|---|---|---|
| Hugging Face | weights | MIT |
| DeepSeek Platform | hosted-api | API key required |
| OpenRouter | hosted-api | API key required |
| Ollama | local | Distilled variants available |
Official sources
- Model card model
- Technical report github
Change log
- — Initial entry.
Esc