Alibaba Unveils Qwen3.8-Flash AI Model, Pioneering Cost-Effective Solutions
On August 26, Alibaba announced the launch of Qwen3.8-Flash, an innovative artificial intelligence model engineered for enhanced cost efficiency. This revelation was shared via a post on social platform X.
Meet Qwen3.8-Flash, a multimodal MoE and an early preview of the Qwen4 architecture, now open-weight!
The production version Qwen3.8-Flash will be available soon via QwenCloud API at just $ 0.16/1M input tokens and $ 0.47/1M output tokens.
125B parameters + 51B N-gram… pic.twitter.com/SScnmzWS7O
— Qwen (@Alibaba_Qwen) August 26, 2026
The Qwen3.8-Flash operates as a multimodal mixture-of-experts (MoE) system, poised for release through the QwenCloud API, featuring a pricing structure of 16 cents per 1 million input tokens and 47 cents per 1 million output tokens.
This model showcases an impressive architecture with 125 billion parameters and 51 billion N-gram embeddings, employing 6 billion activated parameters per token, as detailed in the announcement.
Alibaba emphasized the transformative nature of Qwen3.8-Flash, stating, “Dramatically lower training and inference costs: trained at just 1/9 the cost of Qwen3.7-Plus, while outperforming it across the board with especially strong gains in coding and office tasks.”
Recent reports indicated that Alibaba’s open-weight AI models have ascended to the forefront of global AI offerings, amassing over 3 billion downloads within the past six months—far surpassing competitors such as Meta, Google, and DeepSeek.
As part of its ambitious AI initiative, Alibaba has open-sourced an impressive collection of more than 460 AI models, within which its ecosystem has generated over 300,000 derivatives.
Analyses from open-source AI platform Hugging Face revealed that Google’s offerings yielded 418 million downloads for the year, while Meta’s total stood at 227 million.
Furthermore, on August 14, it was disclosed that Apple collaborated with Alibaba Group to create and train a bespoke AI model tailored for the Chinese market.
This partnership marks a strategic shift for Apple, which had relied on third-party models to fuel AI capabilities in its devices sold within China.
Following this development, Apple is anticipated to unveil its Apple Intelligence suite of AI tools in China in the near future.
Securing a proprietary model developed specifically for China would empower Apple with greater control over the AI features offered in the region, marking a significant achievement as it would become the inaugural foreign entity to gain approval from the Chinese government for a proprietary model.
On August 20, Alibaba disclosed that its AI-related product revenue had achieved remarkable triple-digit growth for the 12th consecutive quarter.
“With our full-stack AI strategy, we have positioned Alibaba to capitalize on the burgeoning demand for artificial intelligence and AI computing,” remarked Alibaba CEO Eddie Wu in an earnings press release.
Source link: Pymnts.com.
