The trend of price reductions and Open Source in large models accelerates, promoting the comprehensive development of AI applications

Recently, Alibaba Cloud announced a significant reduction in the API call prices for its Tongyi Qianwen commercial model and Open Source model. Among them, the input price of the main model Qwen-Long has dropped from 0.02 yuan per thousand tokens to 0.5 yuan per million tokens, a decrease of up to 97%. This move, combined with Alibaba's previously released 110 billion parameter Open Source model Qwen1.5-110B, is expected to accelerate the development and implementation of AI applications.

Price reduction and Open Source have become a common strategy among global large model manufacturers, aimed at addressing two major pain points faced by AI application developers: the high cost of large model APIs and the inadequate quality of open source models. This trend is driving the widespread implementation of AI applications.

In fact, Tongyi Qianwen is not the first large model to implement price cuts. Since May, several large model providers have launched low-priced products or announced price reductions:

1. The second generation MoE model developed by Huanfang Quantitative under the Deep Exploration Open Source is priced at only about one percent of a certain well-known large model's API.
2. Zhipu AI has reduced the calling price of its personal version GLM-3Turbo model by 80%.
3. A well-known AI company has launched a new model with comparable performance but at half the price.
4. Multiple companies such as Doubao, Baidu Intelligent Cloud, etc. have launched free or ultra-low-priced model API services.

This wave of price reductions may stem from advancements in large model inference technology and decreasing costs, providing developers with more options and potentially stimulating a surge in AI application development.

In addition to price reductions, Alibaba Tongyi has also launched eight large language models covering parameters ranging from 500 million to 110 billion, as well as multimodal models for vision, audio, and code. These models cover a variety of application scenarios from mobile to enterprise level, providing developers with a wealth of options.

Overall, large model manufacturers are paving the way for the vigorous development of AI applications through a dual approach of price reduction and Open Source. With continuous technological advancements and ongoing cost reductions, we have reason to expect that AI applications will bear fruit in more fields, bringing new opportunities and transformations to various industries.