Written by Liam Moe and Eduardo Baptista
BEIJING (Reuters) – Chinese tech giants Alibaba and Baidu on Tuesday lowered prices for large-scale language models (LLMs) used to power artificial intelligence (AI) products, as price competition intensifies in China's cloud computing sector. significantly reduced.
Alibaba's cloud division announced that it will reduce various prices of its Tongyi Qwen LLM by up to 97%. For example, the company's Qwen-Long model will go from 0.02 yuan to just 0.0005 yuan per 1,000 tokens (data units processed by LLM) after the price cut.
This was quickly followed by Baidu, which announced a few hours later that it would make the Ernie Speed and Ernie Lite models free for all business users.
There has been a price war in China's cloud computing sector in recent months, with Alibaba and Tencent recently cutting prices for their cloud computing services.
Many Chinese cloud vendors are looking to boost sales as China saw a wave of investment in large-scale language models following the hit debut of US-based OpenAI's ChatGPT in late 2022. relies on AI chatbot services.
Price competition in China's cloud computing space is currently hurting the multilingual models that underpin these chatbots, threatening to cut into companies' profit margins.
Baidu's Ernie Lite and Ernie Speed were released in March, and corporate customers paid to use them until Tuesday.
ByteDance announced last week that it would price its flagship Doubao LLM model 99.3% lower than the industry average for business users.
Chinese LLM developers focus on charging companies fees as a way to monetize their LLM investments.
Some are starting to target individual users. Chinese startup Moonshot recently launched a tipping feature where businesses and individual users can pay for priority use of the company's chatbot service.
Baidu is the first company in China to offer LLM products to paying consumers, charging 59 yuan per month for consumers who want to use its cutting-edge Ernie 4 model.
(Reporting by Liam Mo and Eduardo Baptista; Editing by Susan Fenton)
