Text in, speech out (WAV by default). The opposite direction to Gemini 3.5 Transcribe. Its input window is 8,192 tokens, and Gemini 3.8 Flash-Lite TTS is the cheaper sibling. See modality.
Text in, speech out (WAV by default). The opposite direction to Gemini 3.5 Transcribe. Its input window is 8,192 tokens, and Gemini 3.8 Flash-Lite TTS is the cheaper sibling. See modality.