TLDR
- Alibaba’s Qwen released Qwen3.8-Flash, a new multimodal AI model with improved coding and office-task performance
- The model costs roughly one-ninth of the training cost compared to Qwen3.7-Plus
- Qwen3.8-Flash supports a default context window of 262,144 tokens, expandable to 1 million tokens
- Alibaba also open-sourced Qwen3.8-Flash-Next, which previews the architecture for the upcoming Qwen4 model family
- The launch follows Alibaba’s HK$80 billion share sale and Jack Ma buying over HK$600 million in company stock
Alibaba (BABA) released a new AI model on Wednesday called Qwen3.8-Flash through its Qwen AI division. The model is designed to handle coding and office tasks more efficiently than its predecessor.
Alibaba Group Holding Limited, BABA
Compared to the Qwen3.7-Plus model, Qwen3.8-Flash requires about one-ninth of the training cost. Despite the lower cost, Alibaba says it delivers better performance across coding and office-related tasks.
The model is priced at 0.16 USD per million input tokens and 0.47 USD per million output tokens via API. In Chinese yuan terms, that works out to 1 yuan per million input tokens and 3 yuan per million output tokens.
⚡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
Qwen3.8-Flash supports a default context window of 262,144 tokens. That can be expanded up to 1 million tokens, allowing it to handle large documents, long conversations, and research-heavy workflows.
The model is a multimodal Mixture of Experts (MoE) architecture with 125 billion parameters. Alibaba says its performance is competitive with Anthropic’s Opus 4.6 and DeepSeek’s V4-Flash.
Open-Source Weights and the Road to Qwen4
Alongside the commercial release, Alibaba open-sourced the weights for Qwen3.8-Flash-Next on Hugging Face and ModelScope. This lets developers download, run, and modify the model on their own servers.
Qwen3.8-Flash-Next is being positioned as an early preview of the architecture that will underpin the next-generation Qwen4 model family. It is a sign Alibaba is moving quickly toward its next major model release.
On Monday, Alibaba also launched Wan3.0, its latest AI video generation model, with upgraded capabilities. The company has been rolling out new models at a steady pace this month.
Alibaba Goes All-In on AI Funding
The model launches come shortly after Alibaba announced a HK$80 billion equity offering to fund its AI strategy. The raise is aimed at expanding its full-stack AI infrastructure and capabilities.
Co-founder Jack Ma purchased more than HK$600 million worth of the company’s Hong Kong-listed stock in recent days. Chairman Joe Tsai and CEO Eddie Wu have also been buying, publicly backing the company’s AI direction.
AI has become the biggest growth driver for Alibaba as e-commerce revenue growth slows. Its Qwen models are among the most widely used in China, where competition in the AI space is intensifying.
The Qwen3.8-Flash production version is now live on QwenCloud, with API access coming soon.
Stop guessing and start investing with confidence. KnockoutStocks gives you the AI insights, market intelligence, and stock research you need to spot opportunities, cut through the noise, and make smarter investment decisions — all in one powerful platform.
Sign up today and get 50% OFF full access to our premium stock picks.
Simply use coupon code SPECIAL50 at checkout to claim your exclusive discount.