Alibaba Releases Smaller, Cost-Effective Qwen AI Model

Alibaba Group Holding Ltd. released the latest model under its popular Qwen series, a lower-priced platform aimed at driving adoption of its marquee AI offering globally.

The company made the Qwen3.8-Flash downloadable for use on Wednesday and revealed its weights, the values that help AI systems make decisions. The new offering sports 125 billion parameters — a measure of a model’s sophistication — and its performance is competitive to some of the latest releases by rivals such as Anthropic’s Opus 4.6 and DeepSeek’s V4-Flash, Alibaba said in a statement.

Wednesday’s release uses an architecture designed for the next generation of models: the Qwen 4 series. That new system significantly reduces training and inferencing cost, the company said.

Alibaba is China’s biggest spender on AI, pledging more than 380 billion yuan ($57 billion) over three years toward AI development and infrastructure building. To fund that effort, it raised HK$80 billion ($10.2 billion) this week in Hong Kong’s biggest follow-on share offering, reaffirming a determination to amass and spend vast sums to take global AI leadership.

The e-commerce giant-turned-AI player has ramped up the pace at which it’s deepening and broadening its AI portfolio. Alibaba upgraded its flagship offering Qwen3.8 Max earlier this month, enhancing its largest-ever AI model. It also released the international version of its enterprise-targeting AI agentic tool QwenWork on Wednesday, hoping to expand sales in global markets with a focus on Asia, Latin America and the Middle East.