GLM-5.3 Flash’s Active Parameters, Coinbase’s Tokenized Equities

Z.ai’s GLM-5.3 Flash introduction centers on a model designed for performance and cost efficiency, while Coinbase’s tokenized equities on Base extend stock exposure through crypto infrastructure for non-US users.

Model Scale Versus Active Parameters

Z.ai has introduced GLM-5.3 Flash, a mixture-of-experts model with 320 billion total parameters and 18 billion active parameters. That distinction gives engineers a more specific description of its architecture than the total parameter count alone. Z.ai positions the model as combining high performance with cost efficiency, making it a potential candidate for applications requiring advanced natural language processing. Those are design aims, however, rather than an established cost or performance advantage.

Tokenized Stock Exposure on Base

Coinbase has launched four tokenized equities on Base for non-US users, covering Nvidia, Meta, Apple and Google. The offering brings exposure to those stocks into a cryptocurrency format. For engineers working in fintech, its concrete relevance is the use of Base to deliver products tied to traditional equities. The non-US availability is a material part of the launch’s scope, particularly for a San Francisco audience.