GLM 5.2: Chinese AI Model Disrupts Coding Costs

A new Chinese AI model called GLM 5.2 offers a significantly cheaper option for coding tasks. This model, released by Beijing-based lab Z.ai, costs less than a fifth of the comparable price for access to Anthropic's Opus 4.8 model. It's also a tenth the price of Anthropic's Fable coding model. What's interesting is that GLM 5.2 is an open-weights model, meaning organizations with sufficient hardware can host it for free. For those who pay Z.ai, the API costs $4.40 per million output tokens. This model has 753 billion parameters, with 40 billion active at once, which improves its response speed. The release of GLM 5.2 has raised concerns that U.S. AI companies could lose their competitive edge. It nearly ties Opus 4.8's score on some coding benchmarks and performs well in cybersecurity. Many software engineers, however, are not yet fully aware of the total AI price tag for coding projects and often choose the most powerful model regardless of cost. The bottom line: This new, more affordable Chinese AI model could change how companies approach coding tasks and impact the global AI landscape.