← Catalog
DeepSeek-R1
AvailableDeepSeekFrontierOpen source
Breakout open reasoning model trained with large-scale reinforcement learning and released with weights under MIT.
Specifications
- License
- Open source · MIT
- Weights
- Downloadable
- Architecture
- Mixture-of-Experts
- Parameters
- 671B · 37B active
- Context window
- 128K tokens
- Max output
- 33K tokens
- Knowledge cutoff
- —
- Price (in / out, $/M)
- —
- Modalities
- TextCode
Benchmarks
Vendor-reported figures are claims until independently verified. See methodology.