Alibaba Cloud: Tongyi Qianwen Visual Understanding Model Reduced by 80%

Alibaba Cloud suddenly released a major announcement, announcing a price reduction for large models, with the entire line of Tongyi Qianwen visual understanding models reduced by over 80%. This is also the third round of price reductions following two previous reductions in May and September 2024.

The person in charge of Alibaba Cloud stated, “This price reduction is mainly due to the continuous optimization of Alibaba Cloud’s infrastructure and model structure, as well as the scale effect brought by the exponential growth in large model calls. We hope to release the technical dividend to all users at the first time.” Specifically, Qwen-VL-Plus has been reduced by 81%, with an input price of only 0.0015 yuan per thousand tokens, setting a record for the lowest price across the network; The higher-performance Qwen-VL-Max has dropped to 0.003 yuan per thousand tokens, a decrease of up to 85%. According to the latest prices, one yuan can process approximately 600 images in 720P or 1700 images in 480P at most.

Alibaba Cloud introduced that Qwen-VL is the fastest-growing model on the Alibaba Cloud. It can recognize images of different resolutions and aspect ratios, understand videos longer than 20 minutes, and possess visual intelligence capabilities to operate smartphones and robots independently. It can be applied in various terminal visual recognition scenarios such as mobile phones and cars.

This is also the third round of price reductions by Alibaba Cloud following two price cuts in May and September 2024. In May last year, Alibaba Cloud reduced the price of its Thousand Questions GPT-4 level flagship model Qwen-Long for the first time this year. The API input price of the model decreased from the original 0.02 yuan per thousand tokens to 0.0005 yuan per thousand tokens, a decrease of 97%.

SEE ALSO: Tongyi Qianwen’s Flagship Model Qwen-Long Announces A 97% Price Reduction

At the Apsara Conference in September of that same year, Alibaba Cloud announced another price reduction for three flagship models on the Hundred Refinements Platform. Among them, Qwen-Turbo’s price was directly reduced by 85%, down to 0.3 yuan per million tokens.

Since 2024, there have been waves of price reductions in the field of large models, with technology giants joining this ‘price war’. Not only Alibaba Cloud, but also companies like ByteDance, Baidu (09888.HK), Tencent Cloud (00700.HK), iFlytek (002230.SZ), DeepSeek and Zhipu AI are continuously competing in terms of price and technology.

Regarding the significant price reductions by domestic large model manufacturers one after another, angel investor and artificial intelligence expert Guo Tao stated that domestic general large models have entered a stage where they exchange market share for prices to enhance their market competitiveness. With the continuous development and maturation of technology, the research and development costs of large models are gradually decreasing, enabling manufacturers to lower prices while ensuring certain profits. Furthermore, due to increasingly fierce market competition, major manufacturers are adopting price reduction strategies to compete for market share and attract more customers. In addition, policy support and market demand are also important factors driving the price reduction of large models.

Report