Command Code has made GLM-5.3 FlashX available across its plans and API, the company said in a post on X. The model is described as delivering about 200 tokens per second with a 1 million-token context window.
The release gives Command Code users a model that accepts text, images and video. Its model page lists a 1 million-token context window and prices of $0.37 per million input tokens and $1.25 per million output tokens.
Z.ai, the model's developer, documents inference speeds of 200 tokens per second, a 1 million-token context length and multimodal input. Its documentation also lists 320 billion total parameters, with 18 billion activated parameters, and a maximum output length of 128,000 tokens.
Pandaily reported that Zhipu, also known as Z.ai, runs GLM-5.3-FlashX near 200 tokens per second across roughly 100,000 domestic accelerators. That report describes the model developer's inference deployment.
Command Code's public changelog previously recorded the base GLM-5.3 Flash model in CLI v1.35.0 and Desktop v0.1.19 on Aug. 26, 2026; no separate dated changelog entry for the FlashX variant was located. The base model entry cited image support and a one-million-token context window.
Command Code was founded by Ahmad Awais, its Founder and CEO, and is headquartered in San Francisco. The company describes its product as a coding agent that runs in the terminal and learns a developer's coding style, naming conventions and architectural preferences through its taste-1 system. Its launch page says more than 100,000 developers currently use the product.
