The Chinese start-up Z.ai has unveiled a new artificial intelligence model, GLM-5.2, which has already caught the attention of the tech community. In terms of performance, it is on a par with models from OpenAI and Anthropic, but is significantly cheaper.
- Z.ai has released GLM-5.2, which competes with OpenAI and Anthropic.
- The cost of using the GLM-5.2 is six times lower than that of its American equivalents.
- The model is rapidly gaining popularity amongst developers in the US and Europe.
- Data security remains the main obstacle to implementation.
- GLM-5.2 could significantly intensify competition in the AI market.
The Chinese start-up Z.ai and its new model
The Chinese start-up Z.ai (formerly known as Zhipu AI) has unveiled a new artificial intelligence model GLM-5.2, which has already caught the attention of the Silicon Valley tech community. According to experts, the new product demonstrates a level of performance close to that of OpenAI’s top models and Anthropic, but it operates at a significantly lower cost. Some experts are already calling the launch of GLM-5.2 the «mini-DeepSeek moment» — drawing a parallel with the breakthrough achieved by the Chinese DeepSeek model, which shook up the global artificial intelligence market last year.
Price and features of the GLM-5.2
GLM-5.2 received particularly high marks for its ability to programme and carry out complex tasks with minimal user input. According to a former adviser to the US President on artificial intelligence David Sachs, the new Chinese model is only slightly inferior to Anthropic’s Claude Opus 4.8 and is on a par with GPT-5.5 from OpenAI. On the independent Artificial Analysis platform, GLM-5.2 is already ranked among the world’s top five large language models, and in the Code Arena rankings it occupies second place amongst models for creating websites and front-end applications.
The benefits of open source
At the same time, using the GLM-5.2 costs around six times less than comparable proprietary models from US companies. Experts point out that one of the main advantages of the GLM-5.2 is the model’s open-source nature. It can be deployed quickly without complex fine-tuning, which significantly simplifies integration for companies. The popularity of GLM-5.2 is growing rapidly on the OpenRouter platform, where it has already overtaken some Anthropic models in terms of usage.
Challenges for American companies
Despite technical advances, the widespread adoption of GLM-5.2 in the US and Europe may be held back by mistrust of Chinese technology. Data security is the main cause for concern, particularly amongst banks, government bodies and companies in the cybersecurity sector. At the same time, analysts point out that running the model on in-house servers or via US cloud services significantly reduces these risks.
Competition in the artificial intelligence market
According to the RAND research centre, following the launch of DeepSeek, the global share of Chinese large language models rose from 3% to 13% in just two months. They are used most actively by developing countries, as well as by states that maintain close economic ties with China. Analysts believe that GLM-5.2 is unlikely to displace OpenAI or Anthropic in the near future, but it could significantly intensify competition amongst open-source AI models and force American companies to work harder to improve their own products.
