Today, the Gemini API introduces two new service tiers: Flex and Priority, enhancing cost and reliability control through a single interface. As AI evolves into more complex agents, developers need to manage background tasks, such as data enrichment, alongside interactive tasks like chatbots. Previously, this required splitting architecture between synchronous and asynchronous processes. Flex and Priority streamline this by allowing background jobs to utilize Flex and interactive requests to route through Priority, all via standard synchronous endpoints.
Flex Inference is a cost-effective option that offers up to 50% savings compared to the Standard API by downgrading request criticality. It also simplifies the process by eliminating the need for batch processing and allowing the use of familiar endpoints. Ideal for background CRM updates and large-scale research simulations, Flex provides a seamless way for developers to optimize efficiency while maintaining reliability. Start easily by configuring the service_tier parameter in your API requests.
