Alibaba Claims Its New AI Model Beats Meta's Llama 3, Promising Major Advancements

Alibaba Group Holding Limited BABA, heavily investing in generative AI, has updated its open-source AI models, claiming they surpass Meta Platforms Inc’s META Llama 3 in specific tasks.

Alibaba Cloud launched Qwen2, the second version of its Tongyi Qianwen large language models (LLMs), featuring multilingual pre-training and an expanded context window for longer queries and answers.

Qwen2, which comes in five variations, consistently outperformed Meta’s Llama 3-70B model in benchmarks across various disciplines, the SCMP reports.

Also Read: Temu’s Growth Surge Makes Parent PDD Surpass Alibaba as China’s Most Valuable E-Commerce Giant

The company unveiled Qwen2 shortly after introducing the closed-source Tongyi Qianwen 2.5, which was noted for its superior Chinese language capabilities compared to OpenAI’s GPT-4.

Tencent Holding Ltd TCEHY recently launched Yuanbao, a chatbot powered by its Hunyuan LLM.

Alibaba.com, the company’s B2B cross-border sourcing platform, has implemented AI-powered tools to connect sellers and buyers, according to platform president Zhang Kuo.

Alibaba is also expanding its global cloud computing availability in Mexico and eying new data centers in key markets, including Malaysia, Thailand, and South Korea, over the next three years.

Alibaba Cloud eyes double-digit growth in the second half of 2024.

Alibaba stock has lost 9% in the last 12 months as it combats cutthroat rivalry and a weak economy. Investors can gain exposure to the stock via Avantis Emerging Markets Equity ETF AVEM and Global X Artificial Intelligence & Technology ETF AIQ.

Price Action: BABA shares were trading higher by 1.27% at $79.41 at the last check Monday.

Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.

Photo by Fooksou Lamimo via Wikimedia Commons