DeepSeek-R1-Lite

Released

Early preview of reasoning capabilities with chain-of-thought

Released on 2024.11.20

Overview

DeepSeek-R1-Lite Preview is an early preview of DeepSeek's reasoning model, showcasing chain-of-thought reasoning capabilities. It demonstrates the potential of reinforcement learning for improving model reasoning.

Key Features

  • Chain-of-thought reasoning
  • Preview of R1 capabilities
  • Reinforcement learning enhanced

Specifications

Architecture
Reasoning Model
License
DeepSeek License

Resources