I’m experimenting with various TTS models. Google Cloud Platform’s TTS chirp HD is free for the first million tokens, but the quality is poor, even with a very optimized text.

On the other hand, GPT-4o-mini-tts produces exceptional results. It delivers consistent output for the same text. I’m curious about the absence of an official Gemini TTS. Google Cloud Platform’s chirp HD is a couple of years old.

While ElevenLabs undoubtedly offers the best TTS, I’m hesitant to subscribe to another service for this specific use case.

If you use a TTS model, what are you using currently? Looking for the cheapest option that gives a decent output.