GLM 4.5 is the latest flagship foundation model, purpose-built for agent-driven applications. Built on a Mixture-of-Experts (MoE) architecture, it supports context lengths of up to 128K tokens and offers major improvements in reasoning, code generation, and agent alignment. The model features a hybrid inference system with two modes: a “thinking mode” for complex reasoning and tool use, and a “non-thinking mode” for fast, real-time responses. Reasoning behavior can be easily managed with a simple boolean toggle.
| Creator | zAI |
| Release Date | July, 2025 |
| License | MIT |
| Context Window | 131,072 |
| Image Input Support | No |
| Open Source (Weights) | Yes |
| Parameters | 355B, 32B active at inference time |
| Model Weights | Click here |
