GLM-5.3 pricing and specs

Input costs $1.4 per 1M tokens (118 ₽) and output costs $4.4 per 1M (372 ₽). Zhipu AI bills the prompt and the completion separately, so the price of a call depends on how long both are. Ruble figures use the Bank of Russia rate of 84.51 ₽ per dollar for September 2026.

List prices are quoted per 1M tokens, so 1,000 tokens cost a thousandth of that: divide $1.4 for input and $4.4 for output by 1,000. The same rates in rubles are 118 ₽ and 372 ₽ per 1M tokens. Both prompt tokens and completion tokens are billed.

The GLM-5.3 context window is 1M tokens (1,000,000). Everything shares that budget: the system prompt, the dialogue history, attached files and the answer the model writes. The figure comes from the Zhipu AI model card as of September 2026.

The context window is 1M tokens (1,000,000) and a single response is capped at 131K tokens, so longer output has to be generated in parts. Rate limits are set by Zhipu AI per account and depend on your plan rather than on the model, so they are not listed here.

Providers quote their tariffs in US dollars, so the ruble amounts here are a conversion at the official Bank of Russia rate of 84.51 ₽ per dollar. For GLM-5.3 that gives 118 ₽ per 1M input tokens and 372 ₽ per 1M output tokens. When the rate moves, the ruble price moves with it.

GLM-5.3 belongs to the GLM family at Zhipu AI. Models in the GLM line differ in context size, price per 1M tokens and supported capabilities, so compare them by the numbers on their cards rather than by the name. Every GLM model is listed on the provider page.

Yes, GLM-5.3 declares a reasoning mode: it works through intermediate steps before it answers. Those steps add output tokens, so a call with a long reasoning chain costs more and takes longer than a plain completion. Capabilities declared for the model: reasoning, tool calling, structured output.

The declared capabilities of GLM-5.3 are: reasoning, tool calling, structured output. It accepts text as input, and every attachment consumes tokens from the shared context window of 1M tokens. The list comes from the Zhipu AI model card and is refreshed with the catalog as of September 2026.

Yes, GLM-5.3 ships with open weights, so it can be downloaded and served on your own hardware instead of being used only through the Zhipu AI API. What you may do with it is set by the license shown on the card. Catalog prices cover hosted API access.