Google announced the update of Gemini 2.5 Flash and Gemini 2.5 Pro models for voice synthesis, now available to developers through the Gemini API in Google AI Studio. These models are designed for applications where expressiveness of speech is important, such as audiobook narration, educational courses, product instructions, podcasts, and multi-voice projects.
The update added a wider range of emotional styles and tones, more precise adherence to stylistic cues, intelligent reading speed adjustment depending on context, and more stable multi-voice support, now covering 24 languages. The models replaced previous versions so that users immediately gain access to more natural speech synthesis.
Gemini 2.5 Flash TTS is optimized for quick interactive solutions and is suitable for applications where response time is critical. Gemini 2.5 Pro TTS provides high voice quality, which is important for projects with high sound requirements. Users can finely control speed, tone, and character identity, and the update improved multilingualism.
Partners are already using these models for advanced features, including precise dialogue tuning and pronunciation or intonation adjustments. Early users noted the ability to create cinematic voiceovers for different characters and languages.
Google provides these tools to developers worldwide through Google AI Studio to support the need for creating more realistic and flexible speech synthesis for creative and technical tasks.
