In February 2026, Alex Amederen developed a flexible billing capability for inference tokens in the ubicloud/ubicloud repository, focusing on backend development and billing system design using YAML. Alex introduced a general zero-rate billing rate within InferenceRouterModels, enabling the platform to support out-of-band and rapidly evolving AI models. This approach allowed the billing system to adapt to diverse resource types and usage patterns, aligning pricing mechanisms with the product strategy for AI inference workloads. The work addressed the need for agile cost governance and experimentation, resulting in a robust, adaptable billing infrastructure without reported bugs during the development period.