China's Open AI Models vs. OpenAI and Anthropic: What DeepSeek and Qwen Are Offering

In July 2026, Chinese AI labs led by DeepSeek and Alibaba, with its Qwen family, are pushing their open-source models ever more insistently as a working replacement for the closed flagships from OpenAI and Anthropic. The pitch comes as access to Western frontier models gets more expensive and more restrictive, Wired reports.

What's wrong with closed models?

Access to OpenAI's and Anthropic's most advanced models has become noticeably more limited in 2026: prices are higher, rate limits are tighter, and there are more conditions tied to region and moderation. Wired's assessment is that this is what pushed developers to look for alternatives they can download, deploy on their own servers, and use without depending on someone else's API and its policies.

For many teams, a closed model behind a subscription is also a constant risk: the vendor can raise prices, change usage rules, restrict access by geography, or pull the model altogether. Open weights remove that risk. A model that has already been downloaded cannot be "recalled," which undermines the very business logic of subscription-based access.

What are the Chinese labs offering?

Chinese labs are betting on open weights and permissive licenses. DeepSeek, Alibaba (Qwen), Moonshot AI (Kimi) and Zhipu AI (GLM) release models that can be run locally, fine-tuned for a specific task and built into a product without calling an outside service. Some of them ship under permissive licenses such as Apache 2.0, which allows use in commercial products.

Wired summarizes their offer this way:

Chinese labs are promoting their open-source alternatives as "stable, accessible and increasingly capable." —

Wired

• DeepSeek — a lab based in Hangzhou, known for open reasoning models
• Qwen — Alibaba's family of open models
• Kimi — Moonshot AI's models, with an emphasis on long context
• GLM — Zhipu AI's open models
• The shared selling point — open weights and licenses that permit commercial use

The key difference from the Silicon Valley approach: the model is handed over in full rather than sold through a paid API. It can be inspected, forked and kept entirely under your own control.

Why does this matter for developers?

Open models remove dependence on a single provider: the team controls cost, data privacy and the stability of access. Data does not leave for someone else's cloud, and inference cost is set by your own hardware, not by a vendor's price list.

For companies outside the US, it is also a question of technological sovereignty — not being cut off from a tool because of sanctions, export restrictions or a sudden change in a supplier's policy. According to Wired, in 2026 the main argument for Chinese open models is becoming the predictability of access, not just price. And as these models become "increasingly capable," the quality gap with closed flagships is narrowing as well.

What does this mean?

The balance in the industry is shifting from closed APIs to open weights. Silicon Valley's playbook was built for years on proprietary models sold by subscription; Chinese labs are demonstrating the opposite strategy, in which openness itself becomes a way to capture the market and standardize development around their models. For OpenAI and Anthropic, this means that the closed nature of their frontier models is gradually turning from an advantage into a vulnerability.