Alibaba takes AI fight to Meta with 2 open-source models
Alibaba's cloud computing unit, Alibaba Cloud, has launched two open-source AI models, Qwen-7B and Qwen-7B-Chat, each with seven billion parameters. This marks a significant move for the Chinese tech giant, following Meta's release of the open-source Llama 2 model last month. Analysts believe these open-source models could challenge the market dominance of OpenAI and Google, which charge hefty fees for their AI models.
Alibaba will require licenses from companies with over 100mn users
Alibaba said the two models' "code, model weights, and documentation" will be freely available to academics, researchers, and commercial institutions. Companies with over 100 million monthly active users must obtain a license from Alibaba to use the two new models. Meta's Llama 2 also requires a license when a company has over 700 million users. Llama 2's biggest version, however, has 70 billion parameters.
China is racing to bridge the AI divide
China is striving to close the AI gap with the US, as Beijing pushes for homegrown AI models to compete with American counterparts. Tech giants like Tencent Holdings and Huawei have also been actively developing their own AI models. Alibaba Cloud introduced its large language model, Tongyi Qianwen, in April, offering multiple versions with varying parameter counts.