Revolut says it cut AI coding inference costs by ~90% in less than a year by making self-hosted GLM models its default over commercial tools like Claude Code (Tobias Mann/The Register)

Revolut has reportedly reduced its AI coding inference costs by approximately 90% in under a year. This significant cost reduction was achieved by transitioning to self-hosted GLM models as its default, moving away from commercial tools such as Claude Code. The company found that the availability of cheaper tokens encouraged engineers to engage in time-saving activities like debugging and addressing Jira tickets.