The geopolitical competition between the US and China on the world stage has now become highly intense in the field of Artificial Intelligence (AI). Recently, the US government imposed a sudden ban on foreigners using America's advanced Mythos 5 and Fable 5 AI models. The main reason for this is the fear that Chinese organizations might use these models for cyberattacks.
However, in a way that breaks this US ban, China's leading AI company, Z.ai, has surprised the world by releasing its new GLM-5.2 model. Even in a situation where America's Nvidia chips are not available, this model has been developed using only domestic Huawei chips.
Regarding this, Elon Musk predicted that it would take until 2027 for China to reach this quality. Denying that, Z.ai Chairman Tang Jie has challenged that China will introduce its own 'Mythos-class' AI model equivalent to the US before the end of 2026. This rapid growth is shaping up to reshape the global technological balance.
Z.ai Chairman Challenges Elon Musk's Prediction!
Recently, Z.ai company released its new open-weight model called GLM-5.2. Following this, an interesting discussion arose on the social media platform X (Twitter).
A user posted, "China will create its own AI model equivalent to America's Mythos by November-December 2026." Replying to this, tech billionaire Elon Musk predicted that it might take until the first quarter of 2027 (Q1 2027) for China to reach this goal.
However, Tang Jie, the founder of Z.ai company, has completely denied this prediction by Elon Musk. He hit back strongly, stating, "China will not take that much time to achieve that goal; it will achieve a quality equivalent to American models very soon."
The Monstrous Growth of GLM-5.2 Model: What are the Key Features?
The GLM-5.2 model from Z.ai company, released on June 16, has proved that China can independently create a world-class AI despite US restrictions. Its main features are:
- Massive Parameters: This is a massive open-weight model with 753 billion parameters. It is specifically designed for complex coding and engineering tasks.
- Direct Challenge to US Models: In many key performance tests (Benchmarks), it has shown better performance than OpenAI's GPT-5.5. Furthermore, it works on par with Anthropic's Claude Opus 4.8 model and has outperformed Claude Fable 5 in website design (HTML web design) tests.
- 100% Indigenous Hardware: The US has banned supplying advanced Nvidia chips to China. Because of this, this GLM-5.2 model has been completely and successfully trained using only China's own product, the Huawei Ascend chips. This shows China's hardware self-reliance.
- Cost-Effective API: The API fees for this model, which offers a 1 million token context window, are much lower than Western (US) AI models. Because of this, it is receiving a great response from global companies commercially.
Talent Shortage: The Real Challenge of the AI War!
Even though hardware and technology are on one side, in the current scenario, 38% of global high-quality AI researchers are from China. However, a major setback in this is that only 11% of them stay and work in China.
Most of the remaining Chinese researchers migrate to the US for higher education and employment. Because of this, the US retains 80% of the global AI talent within its own country. Therefore, experts say that not just building supercomputers, but the strategy of retaining these talented people in their own country will be the key factor determining victory in this AI war.
What is its Impact on the Global Market?
This monstrously fast growth of China is going to cause some important changes in the global tech market:
- Failure of US Controls: If China independently creates Mythos-class AI models in a very short period, the export and technology controls imposed by the US will become completely useless.
- Cyber Security Challenges: China developing advanced models with the capability to accurately detect software security vulnerabilities will further intensify the global cyber security race.
- Dominance of Open-source Usage: Since these Chinese models are released under the MIT License, companies and developers around the world can use them in local deployments at a very low cost without any restrictions. This could significantly reduce the market dominance of US companies like Google and OpenAI.
Every ban imposed by the US, instead of paralyzing China's technological growth, is turning into a factor that drives them to grow even faster with self-reliance. Since Nvidia chips were banned, China has succeeded by training world-class models like GLM-5.2 using only its own Huawei Ascend chips.
If Z.ai company breaks the strict US export and technology controls and achieves its goal by the end of 2026, it will stand as a major challenge to America's monopoly on the world stage.
In particular, since these advanced models from China are released in an open-source manner through the 'MIT License', companies around the world can use them without any restrictions. This will significantly reduce the market dominance of US companies like Google and OpenAI. In the battlefield of technological superpowers, it is certain that the global technological balance will completely change by the end of 2026!
Z.ai GLM-5.2 FAQs
1. What is Z.ai company planning to do by the end of 2026? Z.ai plans to develop a highly powerful Artificial Intelligence (AI) model equivalent to America's advanced 'Anthropic Mythos 5' model by December 2026 using domestic technology.
2. What is the special feature of China's GLM-5.2 model? This is an open-weight model with 753 billion parameters. Despite US restrictions, it has been trained entirely using China's own 'Huawei Ascend' chips.
3. Why did the US impose a ban on Anthropic models? Since Anthropic's Mythos 5 and Fable 5 models have the capability to break computer security barriers (Jailbreak) and be used for cyberattacks, the US has banned foreign usage for security reasons.
4. What is China's biggest challenge in the global AI sector? Although 38% of high-quality AI researchers in the world are from China, only 11% of them work in China. Since most of them move to the US, retaining these talented people is the biggest challenge for China.
5. What is the impact on the US due to these open-source models from China? Since Chinese models are available for free or at a very low cost through the 'MIT License', global companies' dependency on US companies like OpenAI and Google will reduce. This will affect America's market dominance.