Developed a pricing system enhancement for the Gemini Live 2.5 Flash Native Audio model within the juspay/langfuse repository, focusing on scalable and maintainable pricing configuration. The work introduced support for text, audio, image, and video token types, enabling more granular and modality-specific pricing entries while removing redundant keys for improved clarity. Leveraging TypeScript and JSON, the implementation aligned pricing semantics with Vertex AI requirements, laying a foundation for future model variants. The approach emphasized clear data modeling and backend development practices, reducing configuration friction and supporting easier onboarding for new models without introducing any bug fixes during the period.