Zhipu AI Unveils New Open-Source Model GLM 4.5

Zhipu AI, a Chinese AI company, has introduced its latest open-source model, GLM 4.5. This release is part of the company's ongoing efforts to expand its General Language Model (GLM) series, which aims to make advanced AI technologies more accessible to developers worldwide.

The new GLM 4.5 model is available under the MIT license, allowing for commercial use and offering developers significant flexibility. Among its features, the GLM-Z1 inference model stands out for its speed, achieving up to 200 tokens per second on consumer-grade GPUs, which is notably faster than some competing models.

Zhipu AI has also launched the GLM-Z1-Rumination-32B-0414, a model designed for autonomous AI agents capable of conducting in-depth analysis and handling complex queries. This model represents a step towards more autonomous AI systems.

In addition to these releases, Zhipu AI has established an open-source fund to support large language model projects, committing significant resources to foster the AI industry ecosystem. The company has received backing from major Chinese tech firms and is actively participating in initiatives to support OpenAI API users affected by regional restrictions.