Chinese AI Company DeepSeek Releases New Reasoning AI Model

The Chinese AI startup DeepSeek launched a preview of its newest model yesterday, DeepSeek-R1-Lite-Preview.

According to TechCrunch, DeekSeek’s new AI model competes with OpenAI’ o1, released in September. The intelligence is capable of reasoning, completing tasks, planning, and processing information before answering a user’s query.

DeepSeek-R1-Lite-Preview is web-based and available for everyone for free. The Chinese startup said the new model has been tested on the American Invitational Mathematics Examination (AIME) and MATH benchmarks. The company said that the API and open-source models will be released soon.

🚀 DeepSeek-R1-Lite-Preview is now live: unleashing supercharged reasoning power!

🔍 o1-preview-level performance on AIME & MATH benchmarks.
💡 Transparent thought process in real-time.
🛠️ Open-source models & API coming soon!

🌐 Try it now at https://t.co/v1TFy7LHNy#DeepSeek pic.twitter.com/saslkq4a1s

— DeepSeek (@deepseek_ai) November 20, 2024

Multiple users on X have already tested the chatbot capabilities and shared results. “Nice release! I tested it on a few math, coding, and knowledge tasks. Lots of thoughts on this one,” said one user. “Deep Think is awesome, congratulations! Reading the inner thoughts removes the impression of a black box like OpenAI o1 and is fun to read. I am very impressed,” added another one.

Other users commented on the speed—the bot takes a few seconds to answer questions—but most agree that this model’s preview outperforms OpenAI o1’s preview.

DeepSeek was founded in 2023, backed by the Chinese quantitative hedge fund High-Flyer Capital Management, and is committed to an open-source approach. They have already released multiple advanced AI models, including DeepSeek Coder V2—which beat GPT4 Turbo—and DeepSeek-V2.5.

Alibaba, one of DeepSeek’s Chinese competitors in the AI race, released over 100 open-source AI models in September.