Alibaba prices $660 AI glasses to rival Meta

Alibaba has set the price of its Quark AI Glasses at 4,699 yuan (US$660), with pre-sales starting October 24, 2025 and shipments expected in December.

The Quark AI Glasses, powered by Alibaba’s Qwen language model and Quark AI assistant, mark the company’s entry into the smart glasses market.

The device supports hands-free calling, music streaming, and real-time language translation.

Alibaba, based in Hangzhou, is positioning the Quark AI Glasses as a competitor to Meta’s smart glasses and Xiaomi’s recent AI eyewear.

🔗 Source: CNBC

🧠 Food for thought

Implications, context, and why it matters.

Alibaba’s glasses lack a clear global plan, making market size hard to gauge

Quark AI Glasses debut on Tmall (Alibaba’s marketplace), JD.com (a major Chinese online retailer) and Douyin (China’s version of TikTok), with no overseas plan 1.
Qwen models promise “over 100 languages” 2, yet core integrations like Alipay (mobile payments), Taobao (shopping or price comparison) and Amap (maps or navigation) remain China-only 1. Without confirmed app store access or partners overseas, the $659 price likely ties to the domestic pre-sale.
Impact on Meta depends on international distribution 3 1. Each market needs new payment links, navigation services and regulatory clearance.

Qwen3 APIs give outside developers quick ways to integrate

Alibaba Cloud Model Studio (the company’s platform for building and deploying AI) offers Qwen APIs with pay as you go pricing 2. In Singapore, Qwen-Flash costs $0.05 per million input tokens (tokens are units of text used for billing), and Qwen-Max is about $1.2. Many models include a 1 million token free quota for 90 days. Qwen3 models that power the Quark AI Chat Assistant sit in Model Studio 2 1.
Batch API calls cost 50% of real time in Singapore, and context caching cuts token use for more savings 4 2. Developer teams should watch Model Studio docs for Qwen3 tiers and any new MCP (Model Context Protocol) tools for the chatbot 1.
Early movers who fit use cases to Qwen3’s improved reasoning and human preference alignment can gain ground 2.