DeepSeek
DeepSeek
DeepSeek is the AI research lab that turned the AI cost conversation on its head. Read more →
Latest DeepSeek updates
7 published updates
DeepSeek V4 Preview Release | DeepSeek API Docs
🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.
DeepSeek-V3.2 Release | DeepSeek API Docs
🚀 Launching DeepSeek-V3.2 & DeepSeek-V3.2-Speciale — Reasoning-first models built for agents!
Introducing DeepSeek-V3.2-Exp | DeepSeek API Docs
🚀 Introducing DeepSeek-V3.2-Exp — our latest experimental model!
DeepSeek-R1 Release | DeepSeek API Docs
* ⚡ Performance on par with OpenAI-o1
DeepSeek API Upgrade | DeepSeek API Docs
Now Supporting Chat Prefix Completion, FIM, Function Calling and JSON Output
DeepSeek API introduces Context Caching on Disk, cutting prices by an order of magnitude | DeepSeek API Docs
In large language model API usage, a significant portion of user inputs tends to be repetitive. For instance, user prompts often include repeated references, and in multi-turn conversations, previous content is frequently re-entered.
Latest DeepSeek videos
0 published videos
DeepSeek is the AI research lab that turned the AI cost conversation on its head. Founded by High-Flyer, one of China's largest quantitative hedge funds, the lab released models in early 2025 that rivaled the world's best at a fraction of the compute cost. DeepSeek R1's release caused a brief but dramatic sell-off in AI infrastructure stocks — a sign of just how much it challenged the assumption that frontier AI requires massive resources.
Timeline
DeepSeek founded as an AI research lab by High-Flyer hedge fund.
DeepSeek V2 released — high performance at a fraction of the compute cost.
DeepSeek Coder V2 released, competitive with leading coding models.
DeepSeek V3 and R1 released; R1 goes viral for matching GPT-4 at low cost.
DeepSeek app tops App Store charts globally; stirs AI policy debate.
Founders & leadership
