AI Breakthrough: DeepSeek R1 Rivals OpenAI

Chinese AI Model DeepSeek R1 Rivals OpenAI at Lower Cost

• DeepSeek R1, developed by Chinese researchers, matches or surpasses OpenAI’s o1 model in reasoning capabilities.
• The model was trained using large-scale reinforcement learning without supervised fine-tuning.
• DeepSeek R1 achieved a 79.8% success rate on the AIME 2024 mathematics benchmark.
• The model costs $0.14 per million tokens, significantly cheaper than OpenAI’s $7.50.
• DeepSeek R1 is open source under the MIT license, allowing unrestricted use and modification.

DeepSeek R1 demonstrates impressive reasoning abilities and cost-effectiveness, outperforming OpenAI’s o1 model on benchmarks while being available as an open-source solution at a fraction of the cost.