DeepSeek
Open-weight model family notable for efficient training at frontier scale.
Overview
DeepSeek is an open-weight model family whose published training methods attracted attention for reaching competitive quality at lower reported cost. It solves the cost question in training rather than in inference, which is where most optimisation effort goes. It exists as evidence that architecture and training-schedule choices matter enough to change what a given budget can produce, and its weights and methods are published.
Technology Graph
How DeepSeek connects to the rest of the stack.
Uses and depends on
- Transformer architecture Algorithms DeepSeek based on Transformer High ✓ source
Applied to
- MoE architecture Algorithms MoE used for DeepSeek High ✓ source
Tools
Software that implements or integrates with DeepSeek.
Infrastructure
Alternatives
Comparisons
Side-by-side breakdowns with the alternatives above.