Hi everyone, I am encountering an issue with gemini-2.5-flash where the call drops with a timeout error. Error: Error: Google API request timed out after 5000ms. ... at listOnTimeout Context: My maxDurationSeconds is set to 1300s. I've successfully had 9-minute calls with this setup before. This specific call dropped exactly at the 5:01 mark. It seems like a latency spike caused the model to take longer than 5 seconds to generate a response, triggering the timeout. My question: Is the 5000ms request timeout hardcoded for the Google provider, or can I increase this limit in the Vapi configuration? The 5s window seems too tight for occasional latency spikes on experimental models. Thanks for your help! https://cdn.discordapp.com/attachments/1440439787067539577/1440439787696947292/Snimka_obrazovky_2025-11-18_o_21.33.03.png?ex=69426ad7&is=69411957&hm=709631bc1c79540794d761a987af81279b5b8b998dd3ca2ec413ad4a9411edb8&

Hi Fanfantulipan, If you're interested in using Google's Gemini models with Vapi, there have been several updates worth noting: 1. **Current Gemini Models Available**: You can utilize the

gemini-2.0-flash-lite

model for reduced latency and lower cost, offering a 1 million token context window [here](https://docs.vapi.ai/changelog/2025/3/15). 2. **Paginated Responses**: All Assistant endpoints return paginated responses, providing metadata like

itemsPerPage

,

totalItems

, and

currentPage

, which can help in handling large datasets [learn more](https://docs.vapi.ai/changelog/2025/3/15). 3. **Real-time Features**: For real-time configurations and experimental features, you can use

gemini-2.0-flash-exp

and

gemini-2.0-flash-realtime-exp

[more info](https://docs.vapi.ai/changelog/2025/1/7). These resources should help you effectively integrate and utilize Gemini models with Vapi.