Chinese AI models are gaining ground on Anthropic and OpenAI after Z.ai released GLM-5.2, an open-source system running at roughly one-sixth the cost of US frontier labs. The launch arrived as Washington tightened access to American models.

The timing reshaped the entire competitive picture across the global AI industry in just one week.

How GLM-5.2 is Reshaping the Chinese AI Race

An open-source AI model is a system whose weights can be freely downloaded, fine-tuned, and run on any infrastructure without the original developer’s permission. GLM-5.2 belongs to that category, and its release has triggered the loudest reaction from Silicon Valley since DeepSeek’s debut last year.

The model carries serious technical credentials. Z.ai, formerly known as Zhipu AI, designed GLM-5.2 with 750 billion parameters and a 1-million-token context window.

Furthermore, the system runs entirely on domestic Chinese chips, a critical detail given the ongoing United States export restrictions.

Benchmarks tell the story. GLM-5.2 now sits within a single percentage point of Anthropic’s Opus 4.8 on a closely watched agentic evaluation.

As a result, the gap between Chinese open models and the very top closed US systems has shrunk faster than most industry forecasts had anticipated.

Follow us on X to get the latest news as it happens.

The release timing was anything but accidental. GLM-5.2 launched a day after Anthropic disabled global access to its most advanced models, including Fable 5 and Mythos. Moreover, OpenAI moved to limit access to GPT-5.6 following a separate government request that same week.

Co-founder Tang Jie addressed the contrast directly. He called the Anthropic suspension “deeply regrettable” and said frontier intelligence should not belong to a few people or be subject to sudden rule changes.

Furthermore, his framing positioned Chinese open weights as the safer institutional bet.

Markets responded immediately. Z.ai shares surged more than 30% in Hong Kong trading and now sit up over 800% since debuting in January. JP Morgan projects Z.ai revenue to expand by more than 534% this year, with profitability arriving by 2028.

Why the Chinese AI Push Now Hits Anthropic and OpenAI

The cost advantage is the most damaging factor for US labs. DeepSeek V4 Pro charges $3.48 per million output tokens. Anthropic’s Fable 5 charged $50 for the same output. As a result, enterprise buyers are now openly rethinking their entire AI vendor relationships.

Adoption metrics support the shift. OpenRouter, a popular AI aggregator platform, now shows that Chinese models hold the top four positions among the most widely used systems globally. DeepSeek, MiniMax, Tencent, and Xiaomi have collectively passed every major US frontier provider by token traffic.

The rotation also extends well beyond price. Open-source models can be downloaded, fine-tuned, and run permanently. As a result, neither developers nor governments can revoke access to a system already running on a customer’s own servers, a quality now suddenly more valuable than raw frontier performance.

The competitive picture remains nuanced. DeepSeek itself estimates that Chinese models trail leading US systems by 3 to 6 months in terms of pure capability.

However, that gap matters less when access becomes the primary risk factor, and pricing determines whether production is viable or token economics are prohibitive.

The broader policy backdrop favors the Chinese push. Washington’s restrictions on Anthropic and OpenAI may end up vindicating China’s broader tech self-sufficiency vision, which accelerated after the 2022 Biden administration chip controls landed.

Furthermore, demand for Chinese open models is rising fastest across developing economies worldwide.

Z.ai also plans a dual listing in Shanghai to fund a long-term push toward artificial general intelligence. The next model, GLM-5.5, is expected to launch in August.

Источник: BeInCrypto