Revolut makes Zhipu GLM the default coding model for about 2,000 engineers

Revolut makes Zhipu GLM the default coding model for about 2,000 engineers

N
News Editor
2026-11-03 10:39:01
Revolut, described in the source as Europe’s highest-valued fintech unicorn, has made Zhipu’s GLM the default AI coding model for roughly 2,000 engineers. Claude Code and Codex have been moved into a secondary role and are now mainly reserved for more complex tasks. David Abitbol, head of AI operations at Revolut, said the company had previously set a $300 monthly allowance per person for Claude Code, but the impact on overall efficiency was limited. In April, Revolut began deploying GLM on its own infrastructure and opened unlimited token usage to engineers. The company is now running GLM 5.3 on about 56 Nvidia B200 GPUs, with an all-in cost of around $0.30 per million tokens. Revolut said the resulting bill is only one-tenth of what it previously paid when using Anthropic and OpenAI. Internal company statistics show coding time metrics improved by about 60%, while the cycle from task intake to delivery improved by about 70%. With token limits removed, engineers have also expanded usage beyond complex coding work to internal document search, Jira ticket handling, and finding requirement details in Slack.

Revolut has made Zhipu AI’s GLM the default AI coding model for about 2,000 engineers, according to ChainCatcher. Tools including Claude Code and Codex have been pushed into a backup role and are now mainly used for more complex tasks.

David Abitbol, Revolut’s head of AI operations, said the company had previously given employees a $300 monthly allowance for Claude Code, but the benefit to overall efficiency was limited. In April this year, Revolut started deploying GLM on its own and opened unlimited token access to engineers.

Cost structure and deployment scale

Revolut is currently running GLM 5.3 on about 56 Nvidia B200 GPUs. The company said its all-in cost works out to roughly $0.30 per million tokens, and the related bill is only one-tenth of what it had previously spent using Anthropic and OpenAI.

Internal productivity metrics

According to internal company statistics, engineers’ coding time metrics improved by about 60%. The cycle metric from taking over a task to delivery improved by about 70%.

When token usage was capped, engineers tended to save their allowance for harder tasks. After those limits were removed, they began using AI much more broadly, including for searching internal documents, handling Jira tickets, and locating requirement descriptions in Slack.

This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan.
6400

Disclaimer:

The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.

Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.

Revolut makes Zhipu GLM the default coding model for about 2,000 engineers | Bit.Fan