DeepSeek-R1 available for premium users!
Blog post from Featherless
DeepSeek-R1, now available for premium subscribers on Featherless, is a groundbreaking model that rivals industry leaders like OpenAI, Google, and Meta in performance while utilizing fewer resources and reducing costs. This release is fully open-source and MIT licensed, showcasing DeepSeek's expertise in mixture-of-experts architecture and GPU compute optimization, which were developed under challenging conditions such as hardware sanctions. The success of DeepSeek, which disrupts the vast budgets and resources of major tech companies, is attributed to a year of careful specialization in these critical domains. Users can interact with DeepSeek-R1 via the Phoenix chat on Featherless, integrate it through the Featherless API, and explore its capabilities in the model catalog.