Alibaba launched an AI model designed to run on consumer hardware like laptops and opened the weights of its most powerful model, a move that keeps the fight for open-weight AI inside the hands of giant corporations while ordinary users get told this is freedom. The Chinese tech giant said on Monday that Qwen3.8-27B has “excellent capabilities” in coding, professional work, research and long-horizon agentic tasks, and said it matches the performance of another model that is ten times its size. The company also said it released the weights of Qwen3.8 Max, its most powerful model, meaning it can be freely downloaded and run. The data and methods used to train that model may not be revealed.
Who Holds the Levers
Alibaba’s release sharpens its competition with Meta in the open-weight AI market. Meta wants to lead in open-source AI, and last week it announced plans to open-source its most powerful AI model and launch new ones designed to run on laptops, in a bid to establish itself as the U.S. alternative to Chinese technology, over frontier labs OpenAI and Anthropic, which use a more proprietary approach. The whole contest reads like a corporate relay race, with one giant trying to outmaneuver another while the public gets marketed a choice between brands of control.
Meta was an early player in open-source AI with its Llama family of models, but was quickly outpaced by Chinese players. Its announcement last week included an open-source family of models called Muse Glimmer, designed to run on a laptop. Nick Patience, AI lead at the Futurum Group, said, “Meta's own re-embrace of open weights ... was itself a response to two years of Chinese labs ... taking a large share” of the open-weight market. That’s the language of competition, not liberation. The market decides who gets to shape the tools.
Who Pays for the Race
Alibaba has established itself as the leader in open-weight AI, with other Chinese companies like DeepSeek and Moonshot also strong players. The popularity and success of open-weight models is often determined by how many times they’re downloaded and how many developers are using them to build their own products. Hugging Face, one of the largest repositories from where developers can download open-weight models, said last week that Qwen-based models now account for 151,448 derivatives, times when an open-weight AI model has been downloaded and used. Hugging Face said this is 2.6 times Meta’s total footprint.
Those numbers matter because they show where the labor and attention go. Developers build on these systems, companies harvest the results, and the hierarchy stays intact. The model may be downloadable, but the power to set the terms still sits with the firms that own the infrastructure and decide what to reveal.
Neil Shah, co-founder at Counterpoint Research, said, “The company which can offer the most capable open weights models will move ahead in this race.” He said, “Alibaba aims to become this undisputed leader, outpacing Meta and eyeing the global market ... as a strong alternative to Silicon Valley frontier-grade deployable models.” Shah said the next battleground for AI models is to succeed running on-device rather than from data centers. Industry experts say on-device AI could perform faster and be more secure because it runs on local hardware like a phone or laptop.
What They Call Openness
The release of Qwen3.8-27B suggests Alibaba is expecting that more advanced models will be run on the edge rather than purely in data centers. Patience said Alibaba has built an advantage in several areas, from open-weight models to on-device AI. He said, “Alibaba has made Qwen the most credible non-US model family to build hardware relationships around, in China and in the open-weight developer community globally.”
That’s the real structure here: hardware relationships, developer communities, market share, and corporate leverage. The language of openness covers a system where the biggest players still decide what gets built, what gets shared, and what stays hidden. Alibaba said the weights can be freely downloaded and run, but the data and methods used to train that model may not be revealed. Freedom, apparently, with a locked back room.
The fight between Alibaba and Meta isn’t about handing power down to people. It’s about which giant gets to sit on top of the stack while everyone else plugs in below.