Alibaba Group Holding BABA has made its Qwen3 Embedding series available for developers as open-source artificial intelligence (AI) models become more popular.
The series marks another addition to the company’s line-up of large language models (LLMs), SCMP reported on Friday.
Alibaba’s new models “support over 100 languages, including multiple programming languages, and provide robust multilingual, cross-lingual and code retrieval capabilities.”
Also Read: China Stocks Get Boost As UBS, HSBC Cite Stimulus Hopes, Tariff Truce, Low Valuations
Alibaba said the new series would enable ongoing optimization of the Qwen foundation model, resulting in enhanced training and improved efficiency of its embedding and reranking systems.
Alibaba Cloud founder Wang Jian said AI will evolve faster than expected and be driven by young innovators solving complex problems. He expects AI to develop at a pace beyond people’s imagination in the next five, 10, and even 50 years, citing examples like DeepSeek.
He had said the capabilities of existing LLMs are far above prior expectations.
In May, Alibaba showcased ZEROSEARCH technology, which it said can cut AI training costs by nearly 90% by helping LLMs simulate search behavior without making actual Application Programming Interface calls to search engines during the training process.
Benchmark analyst Fawne Jiang noted Alibaba as a leading beneficiary of accelerated AI adoption in China and a top structural player in the sector.
The global open-source intelligence market size was worth $14.4 billion in 2024. IMARC Group expects the market to reach $58.0 billion by 2033, at a CAGR of 15.93% from 2025-2033, driven by higher cyber security threats, expansion of e-commerce and online platforms channels, and the implementation of AI security solutions to enrich decision-making processes and address risks within the evolving digital environment.
Price Action: Alibaba shares traded lower by 0.85% at $118.91 at publication on Friday.
Read Next:
Photo by Poetra.RH via Shutterstock