Revolut Cuts AI Coding Costs by 90% with Self-Hosted Models

• Revolut reduced AI coding inference costs by approximately 90% in less than a year.
• The company shifted from primarily using commercial services like Claude to self-hosted models.
• David Abitbol, Revolut's ops and infrastructure lead, shared these details at the AI Engineer conference in New York.
• Revolut reported a 60% improvement in performance after making the switch.
• The company now uses GLM (Generative Language Model) for its self-hosted AI models.