← Catalog
DeepSeek-V2
AvailableDeepSeekOpen source
DeepSeek's first major MoE general model with Multi-head Latent Attention and low-cost API positioning.
Specifications
- License
- Open source · DeepSeek License
- Weights
- Downloadable
- Architecture
- Mixture-of-Experts
- Parameters
- 236B · 21B active
- Context window
- 128K tokens
- Max output
- —
- Knowledge cutoff
- —
- Price (in / out, $/M)
- —
- Modalities
- TextCode
Benchmarks
No benchmark scores recorded yet. Spotted some? Submit a correction.
Vendor-reported figures are claims until independently verified. See methodology.