Alibaba Cloud unveils new AI model to support enterprise intelligence transformation

24/04/2023
264

Large Language Model (LLM), Tongyi Qianwen. The new AI model will soon be integrated into all of Alibaba's services to improve the user experience. The company's customers will have access to the model to cost-effectively develop their own AI features.

Alibaba Cloud also announced lower-cost versions of key cloud products, including Elastic Compute Service (ECS) and Object Storage Service(OSS), OSS Reserved Capacity (OSS-RC) and OSS Anywhere Reserved Capacity (OSS-ARC). These versions make computing more accessible and affordable for businesses that want to take full advantage of the opportunities offered by AI.

Integrating Tongyi Qianwen into Alibaba's business operations

Tongyi Qianwen means "seeking an answer by asking thousands of questions. The AI model will be integrated into all business applications in the Alibaba ecosystem, including business communications, intelligent voice support, e-commerce, search engines, navigation and entertainment. The model supports Chinese and English and will be deployed first on DingTalk, Alibaba's workplace platform for collaboration and application development, and Tmal Genie, a provider of smart home applications with IoT support.

Tongyi Qianwen-powered DingTalk is designed to make internal communication more efficient. For example, it can summarize meeting minutes, turn conversations into text, write emails and create business proposals or plans for promotional campaigns based on simple instructions. Users can even instantly create a mini-application on DingTalk by photographing an idea described on paper.

Tongyi Qianwen-powered Tmall Genie will allow for better conversations with users in China. For example, it can invent and tell children's stories, recommend healthy recipes, offer travel tips and recommend background music for workouts.

In addition, Alibaba Cloud will offer customers access to Tongyi Qianwen in the cloud and support them in setting up their own LLM. By fine-tuning Tongyi Qianwen in a secure cloud environment with proprietary data and industry expertise, companies can create AI models that meet their own business needs. As a result, a new wave of customer growth is expected because no resource-intensive and expensive pre-training processes are required to train the basic models. Tongyi Qianwen is now available in beta for business users in China.

Developers will soon have access to Tongyi Qianwen to develop their own AI applications. This will further strengthen the AI software ecosystem in sectors such as logistics, media, finance, manufacturing, energy and retail. Developers in China can now sign up for API access from Tongyi Qianwen in beta.

Multimodal capabilities such as image recognition and text-to-image conversion will soon be added to the Tongyi Qianwen model.

Generative AI based on LLMs is ushering in an unprecedented new phase. In this new AI era, we can create additional value for our customers and their broader communities through our resilient public cloud infrastructure and proven AI capabilities," said Jingren Zhou, CTO of Alibaba Cloud Intelligence.

Tongyi Qianwen is based on Tongyi, Alibaba's proprietary pre-trained model framework that unifies various AI models, including models that can convert text into images and short videos. Last year, Alibaba Cloud launched ModelScope, an open-source Model-as-a-Service (MaaS) platform with hundreds of AI models, including a Tongyi-based text-to-image model for developers and researchers. To date, ModelScope has more than 1 million active users and 800 available models with more than 16 million downloads.

Accessible and affordable computing for businesses and students

Designed for SMEs, the new line of ECS Universal instances offers the same stability at 40% less cost. This solution is suitable for companies with Web applications and Web sites, business office applications and offline data analysis.

OSS-RC allows customers to reserve storage capacity in a specific region for one year. This reduces capacity costs by as much as 50% compared to pay-as-you-go pricing. When customers are not required to store their data in a specific region, they can create an OSS "Anywhere Bucket" to store data in a region chosen by Alibaba Cloud. OSS-ARC can then reserve capacity for objects stored in OSS Anywhere Buckets. This lowers capacity costs by as much as 70% compared to pay-as-you-go pricing.

To make compute resources more accessible to developers in China, Alibaba Cloud has also announced free trial periods of up to three months including its core products, including ECS and PolarDB databases. Alibaba Cloud will also offer 1,000 free training courses in cloud technology, along with about 500 hands-on experiments for developers based on real business scenarios.

Finally, Alibaba announced that its Cloud for Youth program is partnering with the Unesco Chair of AI in Education to make more compute resources available to students in rural areas of China. Launched in 2021, Alibaba's nonprofit Cloud for Youth program helps reduce the digital divide in China by providing cloud computers to schools in underdeveloped areas. The goal is to help young people strengthen their basic technical and digital skills and foster local digital talent.

Recent

Astralift chooses female leadership: Angelique de Vogel and Charlotte Evers mark new phase

First Online Reading Proficiency Test Measures Reading Improvement with Color

HubSpot launches first free tool for AI search optimisation

29% of European employees believe AI can be a better boss than a human

© Dutch Tech On Heels - 2025
Made with
Web Wings