深度求索(DeepSeek),æˆÂ立于2023年,专注于研究世界领先的通用人工智能底层模型与技术,挑战人工智能å‰Â沿性难题。基于自研è®Â练框架ã€Â自建智算集群和万å¡算力çÂ䏵㾼Â,深度求索团队仅用åŠ年时间便已å‘布并开æºÂ多个百亿级å‚数大模型,如DeepSeek-LLM通用大è¯Â言模型ã€ÂDeepSeek-Coder代ç Â大模型,并在2024å¹´1月率先开æºÂ国内首个MoE大模型(DeepSeek-MoE),å„大模型在公开评测榜å•åŠ真实样本外的泛化效果å‡有超越åŒ级别模型的出色表现。和 DeepSeek AI 对è¯Â,轻æÂ¾æŽ¥å…¥ API。
Email : service@deepseek.com
Intelligence level : advanced
Free to use for everyone
ChatGPT-style intelligence
Drag-and-drop interface for easy configuration
Expert-enabled decision-making
Ask AI questions in natural language
Automated data extraction for efficient operations
Generate human-like text and conversations
Upload PDF documents for analysis