Market
Story
DeepSeek Unveils R1 AI Model Matching OpenAI's o1 at 3% Cost, Excels in Technical Tasks
Jan 24, 2025, 12:26 AM
Chinese AI startup DeepSeek has released its new R1 model, an open-source reasoning artificial intelligence that matches or outperforms OpenAI's o1 model in several benchmarks, particularly in mathematics and coding tasks. DeepSeek R1 operates at a fraction of the cost, reportedly only 3% of the price of OpenAI's o1, with pricing at $0.85 per million tokens for input and $2.50 per million tokens for output. The model achieved a top-three ranking in AI benchmarking platform LMarena, securing the number one position in technical domains such as hard prompts, coding, and math. Developed at a cost of only $6 million using fewer high-end chips, DeepSeek R1 is available under the MIT license, allowing free download, local operation, and modification. The model showcases China's rapid progress in AI development and challenges industry giants by offering a cost-effective and open-source alternative.
View original story
Similar markets
Yes • 50%
No • 50%
Yes • 50%
No • 50%
Yes • 50%
No • 50%
No • 50%
Yes • 50%
No • 50%
Yes • 50%
Partnership with Google • 25%
Partnership with Apple • 25%
Partnership with Microsoft • 25%
No major partnership announced • 25%
