English | ็ฎไฝไธญๆ
A curated list of DeepSeek resources, including cutting-edge AI models, developer tools, research papers, and community projects. Maintained by the community, for the community.
- Official Resources
- Models
- Technical Reports
- Tools and Applications
- Prompts
- Tutorials and Documentation
- Community Projects
- Other Awesome Lists
- How to Contribute
- License
- Acknowledgments
- ๐ DeepSeek Website - Official portal for product updates and research
- ๐ป DeepSeek GitHub - Official code repositories
- ๐ค DeepSeek Chat - Live demo of conversational AI
- ๐ฑ DeepSeek App - Mobile application for iOS and Android
- ๐ API Documentation - Official API reference
- ๐ฐ API Pricing - Detailed pricing information for API usage
- ๐ Service Status - Real-time service status monitoring
- ๐ง DeepSeek-LLM - 7B/67B parameter general-purpose LLM
- ๐ DeepSeek-V2 - MoE model with 236B total params (21B activated), 128k context
- โก DeepSeek-V3 - MoE model with 671B total params (37B activated), 128k context
- ๐ฌ DeepSeek-R1 - RL-enhanced model with 60.9% on MATH, 128k context
- ๐ป DeepSeek-Coder: Let the Code Write Itself - Introduces the first version of DeepSeek Coder model
- ๐ง DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence - Presents an open-source MoE code language model
- ๐ ๏ธ DeepSeek-MoE - Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models
- ๐ผ๏ธ DeepSeek-VL - Vision-language model for real-world understanding (1.3B/7B), 4096 context length
- ๐จ Janus - Unified Multimodal Understanding and Generation Models
- ๐ DreamCraft3D - Official Implementation of DreamCraft3D: Hierarchical 3D Generation with Bootstrapped Diffusion Prior
- ๐งฎ DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models - Introduces a 7B math specialist model achieving 51.7% on MATH benchmark
- โ๏ธ ESFT - Expert Specialized Fine-Tuning
- ๐ awesome-deepseek-integration - A curated list of DeepSeek integration resources
- ๐ ๏ธ awesome-deepseek-coder - A curated list of open-source projects related to DeepSeek Coder
- ๐ DeepSeek LLM: Scaling Open-Source Language Models with Longtermism - Introduces the training methodology and innovations of DeepSeek base models
- ๐ DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model - Presents the architecture and training strategies of DeepSeek-V2
- โก DeepSeek-V3 Technical Report - Details the MoE architecture with 671B total parameters
- ๐ฌ DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning - Introduces reinforcement learning approach for reasoning capabilities
- ๐ป DeepSeek-Coder: When the Large Language Model Meets Programming - In-depth exploration of DeepSeek Coder's technical details and performance
- ๐ง DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence - Presents an open-source MoE code language model
- ๐งฎ DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models - Introduces a 7B math specialist model achieving 51.7% on MATH benchmark
- ๐ DeepSeek Prompt Library - Collection of curated prompts and examples
- โก API Quickstart - Get started in 5 minutes
- ๐ค Hugging Face Models - Model hub integration
- ๐งช Colab Notebooks - Free experimentation
- ๐ Official Prompt Library - Curated collection of prompts and examples
- ๐ Quickstart Guide - First steps with DeepSeek
- ๐ Full Documentation - Comprehensive technical reference
- ๐ค Hugging Face Models - Model documentation and demos
- ๐ Awesome LLM - Large Language Models resources
- ๐ Awesome Open LLM - Open-source LLMs
- ๐ Awesome-Chinese-LLM - A curated list of Chinese LLMs, focusing on small-scale, privately deployable models with low training costs
- ๐ก๏ธ Awesome LLM Security - Security best practices
Your contributions are welcome! Please:
- Read our Contribution Guidelines
- Ensure links are relevant and functional
- Add concise descriptions (50-150 characters)
- Follow the Code of Conduct
Special thanks to:
- DeepSeek engineering team for open-source contributions
- @realShineHuang for maintaining this list
- All contributors (view full list)