HONG KONG — The race to outperform DeepSeek in artificial intelligence capabilities and efficiency heated up further on Tuesday as Alibaba released Qwen3, the latest generation of its open-source large language model family.
The e-commerce giant said Qwen3 significantly reduces computing power compared to its peers and was trained on a huge amount of data — 36 trillion tokens, double the amount used for the earlier Qwen2.5 model.
