Google has announced significant upgrades to its Gemini AI suite this week, introducing advanced text-to-speech capabilities, improved Search integration, and expanded global support for preferred sources. The Gemini 2.5 text-to-speech models, including Flash TTS and Pro TTS, now offer developers greater control over voice tone, pacing, and multi-speaker consistency. These improvements enable the creation of interactive virtual assistants, audiobooks, podcasts, and localized audio narratives with more natural and expressive outputs.
Multi-speaker handling is enhanced, allowing consistent voice identities across interviews or scripted dialogue. Developers can customize accents and speaking styles to suit regional preferences or storytelling needs. These models are accessible through the Gemini API within Google AI Studio.
In addition, Google is expanding Search functionalities, including a global rollout of the Preferred Sources feature, enabling users to prioritize content from selected publishers in their feeds. Inline links in AI Mode responses are increased for better content visibility, and subscription links from news providers are now highlighted more prominently. The Web Guide experiment is also expanded to organize links thematically for easier discovery of related content.
Furthermore, Google is piloting commercial partnerships with global and regional media organizations to enhance AI-powered article overviews, audio briefings, and attribution-linked content. Collaborations with real-time information providers aim to improve query results and overall user engagement within the Gemini app and AI Mode in Search. These updates mark a major step in Google’s effort to combine AI innovations with news accessibility and personalized Search experiences.