Revolut Switches to Self-Hosted GLM for Coding Tasks
Fintech giant Revolut has migrated approximately 2,000 engineers from frontier APIs like Claude Code and Codex to a self-hosted GLM model for coding tasks. Leveraging 56 Nvidia B200 GPUs, the setup achieves a blended cost of $0.30 per million tokens—roughly one-tenth of previous API expenses—and reduced engineering lead time by 70%. This shift highlights how open-weight models are becoming the default when cost barriers vanish, posing a significant pricing challenge for proprietary AI labs.
Full Report
Guillermo Flor (@guilleflorvs) · 2026-11-02 22
BREAKING: Revolut moved ~2,000 engineers onto self-hosted GLM for coding and pushed Claude Code and Codex to the exception list.
Per David Abitbol, who leads AI ops there: 56 Nvidia B200s, a blended 30 cents per million tokens, roughly a tenth of what they paid Anthropic and OpenAI. Engineering lead time improved 70%.
Here's the part that matters: the gains showed up when tokens stopped being rationed. Engineers quit saving AI for the hard tickets and pointed it at everything.
Frontier models become the exception, open weights become the default. That's a pricing problem for every lab heading to IPO.
h/t @OliverMolander
Would you move your team off frontier APIs if the bill dropped 90%?
