News
Introducing DeepSeek-V3 2024/12/26
News
⚡ 60 tokens/second (3x faster than V2!)
💪 Enhanced capabilities
🛠 API compatibility intact
🌍 Fully open-source models & papers
Introducing DeepSeek-V3 2024/12/26
🧠 671B MoE parameters
🚀 37B activated parameters
📚 Trained on 14.8T high-quality tokens
Biggest leap forward yet
Model 👉 https://github.com/deepseek-ai/DeepSeek-V3
Paper 👉 https://github.com/deepseek-ai/DeepSeek-V3/blob/main/DeepSeek_V3.pdf
⚡ 60 tokens/second (3x faster than V2!)
🎉 Until Feb 8: same as V2!
🤯 From Feb 8 onwards:
💪 Enhanced capabilities
🌌 Open-source spirit + Longtermism to inclusive AGI
🌟 DeepSeek’s mission is unwavering. We’re thrilled to share our progress with the community and see the gap between open and closed models narrowing.
🚀 This is just the beginning! Look forward to multimodal support and other cutting-edge features in the DeepSeek ecosystem.
💡 Together, let’s push the boundaries of innovation!
Biggest leap forward yet
🎉 What’s new in V3
💰 API Pricing Update
Still the best value in the market! 🔥
