On Tuesday, Alibaba Cloud announced a significant price reduction for its most advanced large language model service, with discounts of up to 85%, aimed at attracting more Chinese enterprise customers. The news was released on the WeChat platform and first reported by the South China Morning Post.
The Qwen-VL-Max model, part of Alibaba Cloud's offerings, is now available at a rate of 0.003 yuan (approximately $0.00041) per thousand tokens. Qwen-VL-Max is a visual reasoning model capable of perceiving and understanding both text and image inputs. The new price is significantly lower than that of similar models from competitors like ByteDance.
Alibaba Cloud, the cloud computing division of the e-commerce giant Alibaba Group, has made this move to highlight the intense competition among major Chinese tech companies in the emerging AI business.
In the Chinese AI market, Alibaba Cloud faces competition from Tencent, Baidu, JD.com, Huawei, and ByteDance (the parent company of TikTok). Over the past 18 months, these companies have all launched competitive large language models, aiming to capitalize on the surge in generative AI technology.
Large language models (LLMs) are AI models trained on vast amounts of data to generate human-like responses to user queries and prompts. They form the foundation of various generative AI chatbots, such as Google’s Gemini and OpenAI’s ChatGPT, as well as next-generation search engines like Perplexity AI and image generators like DALL-E.
Alibaba Cloud primarily targets Chinese enterprises, which, like their American counterparts, are interested in the potential of generative AI to boost productivity. In May, Alibaba Cloud stated that over 90,000 Chinese businesses had downloaded its Qwen model.
However, Alibaba Cloud faces fierce competition. In the past year, Chinese generative AI developers have launched more than 250 new large language models for public use. In addition to the major tech companies, China also has several popular startups, such as DeepSeek, which recently announced its DeepSeek-V3 model with 67.1 billion parameters, one of the most powerful open-source models available.
The Qwen model series from Alibaba Cloud includes Qwen-VL, Qwen-VL-Chat, Qwen-VL-Max, Qwen2-VL, and the experimental QVQ-72B-Preview. Among these, Qwen2-VL-Max has shown excellent performance in benchmark tests such as DocVQA and MathVista, outperforming OpenAI’s GPT-4V and Google’s Gemini Ultra.
This is not the first time Alibaba Cloud has used significant price cuts to attract more business. In February, the company announced price reductions of up to 55% for several of its core cloud computing services, and in May, it reduced the price of its original Qwen-VL model by 97%.
