Grok TTS Scores 96 on Vapi Humanness Index, Tops AI Voices

On June 18, 2026, xAI said its speech-synthesis model "Grok TTS" had scored 96/100 on the "Humanness Index™" — a public benchmark from Vapi that measures the naturalness of AI voices through blind voting — ranking first among AI voice models. The gap to the 100 points assigned to a real human voice is just four points, signaling that AI-generated speech is nearing a level that is hard to distinguish from human speech.