← Back to DeepSeek updates ← Terug naar DeepSeek-updates
DeepSeek ARTICLE ARTIKEL 20 January 2025 20 januari 2025

DeepSeek-R1 Release | DeepSeek API Docs DeepSeek-R1 Release | DeepSeek API Docs

* ⚡ Performance on par with OpenAI-o1 * ⚡ Performance on par with OpenAI-o1

DeepSeek-R1 Release | DeepSeek API Docs
Article details Artikelgegevens
AI maker AI-maker DeepSeek Type Type Article Artikel Published Gepubliceerd 20 January 2025 20 januari 2025 Updates Updates Videos Video's View original article Bekijk origineel artikel

News

DeepSeek-R1 Release 2025/01/20

⚡ Performance on par with OpenAI-o1

News

📖 Fully open-source model & technical report

DeepSeek-R1 Release 2025/01/20

🏆 Code and models are released under the MIT License: Distill & commercialize freely!

⚡ Performance on par with OpenAI-o1

🌐 Website & API are live now! Try DeepThink at chat.deepseek.com today!

⚡ Performance on par with OpenAI-o1

🔥 Bonus: Open-Source Distilled Models!

📖 Fully open-source model & technical report

🔬 Distilled from DeepSeek-R1, 6 small models fully open-sourced

📖 Fully open-source model & technical report

📏 32B & 70B models on par with OpenAI-o1-mini

🏆 Code and models are released under the MIT License: Distill & commercialize freely!

🤝 Empowering the open-source community

🏆 Code and models are released under the MIT License: Distill & commercialize freely!

🌍 Pushing the boundaries of open AI!

🌐 Website & API are live now! Try DeepThink at chat.deepseek.com today!

📜 License Update!

🌐 Website & API are live now! Try DeepThink at chat.deepseek.com today!

🔄 DeepSeek-R1 is now MIT licensed for clear open access

🔥 Bonus: Open-Source Distilled Models!

🔓 Open for the community to leverage model weights & outputs

🔥 Bonus: Open-Source Distilled Models!

🛠️ API outputs can now be used for fine-tuning & distillation

🔬 Distilled from DeepSeek-R1, 6 small models fully open-sourced

🛠️ DeepSeek-R1: Technical Highlights

🔬 Distilled from DeepSeek-R1, 6 small models fully open-sourced

📈 Large-scale RL in post-training

📏 32B & 70B models on par with OpenAI-o1-mini

🏆 Significant performance boost with minimal labeled data

📏 32B & 70B models on par with OpenAI-o1-mini

🔢 Math, code, and reasoning tasks on par with OpenAI-o1

🤝 Empowering the open-source community

📄 More details: https://github.com/deepseek-ai/DeepSeek-R1/blob/main/DeepSeek_R1.pdf

🤝 Empowering the open-source community

🌐 API Access & Pricing

🌍 Pushing the boundaries of open AI!

⚙️ Use DeepSeek-R1 by setting model=deepseek-reasoner

🌍 Pushing the boundaries of open AI!

💰 $0.14 / million input tokens (cache hit)

📜 License Update!

💰 $0.55 / million input tokens (cache miss)

📜 License Update!

💰 $2.19 / million output tokens

🔄 DeepSeek-R1 is now MIT licensed for clear open access

📖 API guide: https://api-docs.deepseek.com/guides/thinking_mode

🔄 DeepSeek-R1 is now MIT licensed for clear open access

More from DeepSeek Meer van DeepSeek

All updates Alle updates

Gemini komt eraan