RT Google DeepMind (@GoogleDeepMind)We released an updated Gemini 2.5 Flash Native Audio model for live voice agents. 💬It's now even better at following instructions and holding more natural conversations. Find out more → https://t.co/nlnLDCQg9I ...