Zhipu AI Launches Free GLM-Image with Strong Chinese Text, Culture Skills

Zhipu AI has launched GLM-Image, a new standalone image generation tool available at image.z.ai. The service is free, provides unlimited use, and produces 2K resolution images without watermarks. It has attracted considerable attention within the Chinese artificial intelligence community for its advanced Chinese language capabilities, according to Pandaily.

Credit: Z.AI

The GLM-Image model demonstrates notable skill in understanding Chinese cultural contexts, a feature which sets it apart from some competitors. Prompts involving classical Chinese poetry references, such as "a solitary smoke straight in the desert, the setting sun round over the long river," consistently result in compositions capturing the artistic essence and original atmosphere. This avoids producing mere collage style illustrations lacking deeper contextual resonance. For scenes depicting spring in a Jiangnan water town, the system capably renders misty rain effects alongside authentic architectural textures, enhancing visual realism.

A key feature of this tool involves its capability for accurate Chinese text generation within images, addressing a common limitation of prior systems. Previous artificial intelligence image tools commonly produced illegible or garbled Chinese characters, rendering them unsuitable for practical applications requiring written communication. GLM-Image, however, achieves precise rendering of multi character Chinese text in natural settings. For instance, tests showed successful generation of promotional posters displaying campaign slogans like "Second Cup Half Price," with all characters correctly formed and seamlessly integrated into the design. But minor character imperfections sometimes appear in more complex compositions, such as calligraphic text on a traditional fan surface, where brushstroke aesthetics are crucial.

The model also performs well in physically rendering transparent and reflective materials, a complex visual challenge. In specific glassware tests, GLM-Image accurately depicts light refraction through liquid filled transparent containers. This demonstrates an advanced understanding of physical optics, a capability that previously often required high end rendering software and considerable computational power.

When drawing comparisons to paid alternatives such as Midjourney and Google's Nano Banana Pro, GLM-Image exhibits slightly less refinement in overall texture detail. Its rendering of long text also occasionally produces minor errors, particularly when complex stylistic requirements are introduced. However, considering its zero cost, unlimited usage, and robust policy of no watermarks, coupled with its native Chinese text capability, the tool provides compelling value for everyday users. These users might frequently be creating promotional materials, social media content, or initial design mockups, where cost and accessibility are paramount. And the organisation Zhipu AI is a major large language model company in China, valued at over ten billion Singaporean dollars.

The GLM-Image tool is broadly accessible via the Zhipu Qingyan platform and its dedicated website, image.z.ai. It does not require a virtual private network, nor does it impose credit quotas on users. Furthermore, it supports natural Chinese language prompts, simplifying user interaction and enhancing its utility for a specific demographic.

• Zhipu AI launched GLM-Image, a free image generation tool.

• It offers unlimited usage and 2K resolution without watermarks.

• The tool excels at accurate Chinese text generation and cultural context understanding.

• It demonstrates advanced physical rendering of transparent and reflective materials.

• GLM-Image is accessible without VPN or credit quotas, supporting natural Chinese prompts.

Source: Pandaily