Best TTS models, a blind benchmark
In 2026, a blind benchmark tested 16 text-to-speech models to determine their quality. xAI and Gemini emerged as the top performers across various scenarios, with xAI excelling in code-switching tests. Groq Orpheus surprised by winning in expressive annotations, showcasing the diversity in TTS capabilities.
- ▪The benchmark included models from ElevenLabs, OpenAI, Azure, AWS Polly, and others.
- ▪xAI and Gemini were the top two models in dialog and narration quality tests.
- ▪Groq Orpheus won in the expressive annotation category, with xAI and Gemini close behind.
Opening excerpt (first ~120 words) tap to expand
Best Text-To-Speech Model in 2026: Blind BenchmarkMay 12, 2026 · 7 min readLewis DwyerSoftware engineer and technical writerText to SpeechIn 2026, the text-to-speech (TTS) market is saturated. Every provider is offering a new groundbreaking model trained on a zillion hours of natural human speech in 200 different languages. Many run marketing pages with blind tests pitting their own model against a competitor's flagship and asking which you prefer. Theirs usually sounds great. The competitor sounds like HAL 9000. To provide an unbiased platform for a real comparison, we tested 16 different models in a blind head-to-head: ElevenLabs, OpenAI, Azure, AWS Polly, Gemini, xAI, Deepgram, Hume, Voxtral, Rime, Cartesia, Groq Orpheus, LMNT, Inworld, Smallest, and UnrealSpeech.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Hacker News (Newest).
Discussion
0 commentsMore from Hacker News (Newest)
