Key Info

Microsoft AI's two text-to-speech models, MAI-Voice-2.1-Flash and MAI-Voice-2.1, are now live on OpenRouter, priced at $15 and $22 per 1M characters respectively. Both can render the same voice across 23 languages while preserving a native accent.

Highlights

  • MAI-Voice-2.1-Flash: optimized for speed, generates 45 seconds of audio within 150ms, targeted at voice agents and call centers
  • MAI-Voice-2.1: highest-fidelity, studio-grade expressive speech for audiobooks, podcasts, and narration
  • Both support 23 languages with a consistent voice and native accent
  • Pricing: $15/1M chars (Flash) and $22/1M chars (standard)