Zhipu AI Unveils GLM-4 Model with Advanced Performance Paralleling GPT-4

Zhipu AI, an emerging player in the AI sector, has unveiled its latest large model, GLM-4, at the “Zhipu DevDay” event in Beijing held on Jan 16. The company claims that GLM-4’s performance closely parallels that of OpenAI’s GPT-4.

The GLM-4 model boasts a range of enhanced features, including support for longer context lengths, improved multimodal capabilities, and faster inference speed. It’s also designed for higher concurrency, aiming to reduce inference costs, thus making it a potentially more economical choice for businesses.

A significant upgrade in GLM-4 is its enhanced intelligent agent capabilities. The model introduces GLM-4 All Tools, a feature that autonomously understands and plans complex instructions based on user intent. This allows it to interact with web browsers, interpret code, and work with multimodal text-generation large models to accomplish complex tasks. 

For instance, GLM-4 can handle a context window length of 128k, and process text equivalent to 300 pages with a single prompt. In the needle test, a measure of information retrieval accuracy, GLM-4 reportedly achieves almost 100% accuracy within a 128K text length. The model also demonstrates enhanced multimodal capabilities, including text generation and multimodal understanding.

The All Tools capability of GLM-4, a new feature, leverages the model’s robust agent abilities. It can autonomously interpret and plan complex instructions, interacting with web browsers, code interpreters, and the CogView3 model of text generation. Furthermore, it can automatically call a code interpreter for solving complex equations or calculus problems. When tested on the GSM8K, Math, and Math23K datasets, GLM-4 All Tools achieved results on par with GPT-4 All Tools.

In addition to its capabilities, the All Tools feature of GLM-4 is fully automatic and versatile, capable of handling a wide range of tasks such as file processing, data analysis, and chart drawing. It can also handle various file formats, including Excel, PDF, and PPT.

Zhipu AI has previously collaborated with the CCF China Computer Society and the Social Media Processing Committee of the China Association for Information Processing. Together, they launched funds to support research and cross-disciplinary innovation related to large models. In 2024, the company plans to launch an open-source large model fund to further contribute to the large model open-source community.

Zhipu AI’s CEO, Zhang Peng, emphasized the company’s commitment to the long-term development and innovation of large models. He expressed hope for collaboration with all stakeholders in the AI industry, including partners in the industrial chain, developer communities, and academia, to contribute to China’s AI future.

The company’s progress has been backed by substantial financial support. In October 2023, Zhipu AI announced it had raised over 2.5 billion yuan in funding, with investments from prominent institutions such as the Social Security Fund Zhongguancun Independent Innovation Fund, Meituan, Alibaba, Tencent, Xiaomi, Kingsoft, Sequoia, Gaorong, and others.

SEE ALSO: OPPO Establishes AI Center to Strengthen Competitive Edge