MiniMax announced MiniMax Speech 2.6, the company’s newest text-to-speech (TTS) / text-to-audio engine optimized for real-time voice agents, voice cloning, and high-fidelity narration. The update focuses on ultra-low latency, smarter handling of technical formats (URLs, phone numbers, dates, amounts), and a new “Fluent LoRA” pipeline to make cloned voices sound natural and fluent across languages. […]