Revolut makes self-hosted GLM its default AI coding model

Revolut has made self-hosted GLM its default for AI coding. Its infrastructure lead puts the inference bill at roughly one-tenth of prior Anthropic and OpenAI spending. He also reported a 60% improvement in coding time and a separate 70% improvement in lead time after AI tools reached roughly 2,000 engineers. Those figures lack published measurement details, and Revolut still uses other models for specialized tasks.