Alibaba Cloud once again announced price adjustments for its large model Tongyi Qianwen series visual understanding models. This is the third price reduction this year, with a price reduction of more than 80%. This move will significantly reduce user costs and further promote the popularization and application of AI technology. This price reduction covers multiple models such as Qwen-VL-Plus and Qwen-VL-Max. Among them, the price of the Qwen-VL-Plus model has reached a new low on the entire network, bringing more application opportunities to developers and enterprises.
Following two price cuts in May and September this year, Alibaba Cloud once again announced price cuts for its large models, launching the third round of price adjustments this year. The price reduction this time is very significant. The price of Tongyi Qianwen series visual understanding models has been reduced by more than 80%.
Among them, the price of the Qwen-VL-Plus model dropped by 81%, with the input price being only 0.0015 yuan/thousand tokens, setting a new lowest price on the entire network; while the higher-performance Qwen-VL-Max dropped in price to 0.003 yuan/thousand tokens, a decrease of 85%. According to the new pricing, 1 yuan can process up to about 600 720P pictures, or 1,700 480P pictures.
The Qwen-VL series large models are multi-modal large models launched by Alibaba Cloud. They have become one of the most popular models in the open source community and have powerful visual reasoning capabilities. This model can not only recognize pictures of different resolutions and aspect ratios, but also understand long videos of more than 20 minutes, and has the visual understanding ability to autonomously operate intelligent objects such as mobile phones and robots. Qwen-VL is widely used in visual recognition scenarios of various terminals, covering mobile phones, automobiles and other fields.
The Alibaba Cloud Bailian team stated that 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 about by the exponential growth of large model calls. With the continuous advancement and optimization of technology, Alibaba Cloud's reasoning efficiency has been greatly improved. The elastic AI computing power scheduling system built by Alibaba Cloud, combined with the Bailian distributed inference acceleration engine, not only greatly reduces the cost of model inference, but also speeds up the inference speed. Alibaba Cloud also mentioned that as the visual understanding effect of Qwen-VL continues to be optimized, this model has become one of the fastest growing models on the Bailian platform.
In order to further reduce the cost for users to use large model APIs, Alibaba Cloud Bailian also launched a new KV Cache billing model. This mode automatically caches context to avoid repeated calculations, thereby significantly reducing model calling costs. It is especially suitable for scenarios such as long text, code completion, multi-round conversations, and specific text summaries.
As Alibaba Cloud continues to optimize infrastructure and models, the price reduction of the Qwen-VL series of visual understanding models not only makes AI technology more accessible to the people, but also brings more application opportunities to developers and enterprises. By continuously optimizing performance and reducing usage costs, Alibaba Cloud is promoting the popularization and application of AI technology and providing stronger technical support for the digital transformation of various industries.
This price reduction reflects Alibaba Cloud's determination to lower the threshold of AI technology and promote inclusive AI. It injects new vitality into the development of the industry and heralds a broader future for AI applications.