Zhipu has open-sourced the GLM-5.3 model weights, making them available for download with support for local deployment and customized use. The company said the model is designed for complex coding, defensive cybersecurity and long-horizon tasks. GLM-5.3 scored 60 points on the AA Composite Intelligence Index, matching closed-source flagship models including Claude Fable 5 and GPT-5.6 Sol. It also tied with Kimi K3 for first place among open-source models. A security review is required only when institutions with annual revenue above $10 billion offer the model as an external service.