Model positioning
GLM-Z1-AirX is the high-speed version of GLM-Z1-Air, positioned as an ultra-fast reasoning model. It was officially released on April 14, 2025, with the release identifier GLM-Z1-AirX-0414. (docs.bigmodel.cn)
Capabilities and uses
The model accepts text input and produces text output, and is designed for complex tasks such as logical reasoning, mathematics, and programming. It is suitable for high-frequency invocation scenarios requiring fast responses, such as real-time data analysis and intelligent customer service. At launch, the official announcement stated that its speed could reach up to 200 tokens/second; this is a maximum speed, not a guaranteed speed for every invocation. (docs.bigmodel.cn)
Limitations and availability
- The context window is 32K, and the maximum output is 30K tokens.
- The official model documentation states that the GLM-Z1 series has been discontinued. Although the pricing table still lists this model, that does not establish that it can currently be invoked.
- Generated results may contain errors; important conclusions and code should be verified by a human. (docs.bigmodel.cn)
Pricing
The official pricing table for Zhipu's domestic platform lists the following prices in RMB: input 5 CNY/million tokens, output 5 CNY/million tokens. These apply to the model's 32K context; cache hits are not supported. These are the prices listed in the table and do not indicate that the model is currently available. Date checked: November 18, 2026. (docs.bigmodel.cn)
The official international site's USD pricing page was also checked, but no USD input or output price for this model was found. The USD price was therefore not verified, and no currency conversion was made. Date checked: November 18, 2026. (docs.z.ai)