Kai-Fu Lee: Yi-Lightning Model Surpasses GPT-4o, China and U.S. Leading Models Only 5 Months Apart
李开复：Yi-Lightning模型超越GPT-4o，中美顶尖模型只差5个月

On October 16, Kai-Fu Lee, founder of Zero-One Manner, announced the launch of a new pre-trained model, Yi-Lightning.
10月16日，零一万物创始人李开复宣布推出新的预训练模型Yi-Lightning。

This model surpassed OpenAI's GPT-4o on the authoritative LMSYS leaderboard, ranking sixth in the world and first in China.
该模型在国际权威榜单LMSYS上超越了OpenAI的GPT-4o，排名世界第六，中国第一。

Yi-Lightning's generation speed has increased by 40%, costing only 0.99 yuan per million tokens, making it suitable for app usage and enterprise applications.
Yi-Lightning的生成速度提高了40%，每百万token仅需0.99元，适合App和企业应用场景。

Lee emphasized that Yi-Lightning is the first Chinese model to surpass GPT-4o on an international leaderboard.
李开复表示，这是中国首个在国际榜单上超越GPT-4o的模型。

Although there is still a gap between Chinese and American models, Lee pointed out that the gap has narrowed to just five months.
尽管中美模型仍有差距，但李开复指出，差距已经缩小至五个月。

Zero-One Manner has completed three rounds of financing, with a valuation exceeding $1 billion, becoming a unicorn company.
零一万物已完成三轮融资，估值超过10亿美元，成为独角兽企业。

逐句精讲 | 李开复：Yi-Lightning模型超越GPT-4o，中美顶尖模型只差5个月 05:01 Share

每日新闻 | 英语精讲

Shownotes Transcript

逐句精讲 | 李开复：Yi-Lightning模型超越GPT-4o，中美顶尖模型只差5个月