Google has recently updated the 2.5 Flash and Flash Lite models and now they are more intelligent, cheaper and faster. This models are available under the preview-09-2025 and Google has created gemini-flash-latest. I would like to know which endpoint is vapi.ai using? And if it is using the old one, when will this models become available to general use on vapi.ai? https://cdn.discordapp.com/attachments/1421451205070160013/1421451205573349386/new-checkpoint-of-gemini-2-5-flash-and-flash-lite-just-v0-37jnfr7lterf1.webp?ex=68d91518&is=68d7c398&hm=c67bfc73df18a314e2da66781302c7d56a030cde35a1f2a6ac3d4717a66708f0&

Currently, Vapi.ai supports the

gemini-2.5-flash-preview-04-17

model for advanced capabilities. There is no documentation confirming support for the "gemini-flash-latest" or "preview-09-2025" endpoints yet, nor a public timeline for their general availability. To check for the latest supported models, review the [Vapi changelog](https://docs.vapi.ai/changelog/2025/4/27) regularly for updates. Source: - [Changelog April 27, 2025](https://docs.vapi.ai/changelog/2025/4/27)