Audio Models
Text-to-speech, voice and music generation models, from expressive narration to radio-ready songs and commercial-safe audio.
Eleven v4
ElevenLabs' September 28, 2026 text-to-speech pair — Eleven v4 for produced, emotive speech and Eleven v4 Turbo for real-time agents at about 100 ms inference. 90+ languages, audio tags, #1 and #2 on the Artificial Analysis Speech Arena, from $0.04 per 1,000 characters.
Gemini 3.8 Flash TTS
Google's Sept 23, 2026 flagship text-to-speech model — voice design from a text prompt, 130 languages, line-by-line direction, two-speaker scenes and long-form stability, at $0.00225 per 10 seconds of audio through 2026.
Suno v5.5
The leading AI music generator for full, radio-ready songs — natural vocals, real song structure, voice cloning and 30+ genres from a single text prompt.
ACE-Step
The open-source full-song generator — free, self-hostable music with vocals that runs on a single consumer GPU.
Amazon Nova 2.5 Sonic
Amazon's Oct 5, 2026 speech-to-speech model for real-time voice agents — better reasoning, instruction following and tool-calling accuracy and lower latency than Nova 2 Sonic, 18 voices in seven languages, 256K context, same price, on Amazon Bedrock.
MAI-Voice-2.1
Microsoft AI's Oct 1, 2026 text-to-speech pair — MAI-Voice-2.1 for expressive long-form audio at $22 per 1M characters and MAI-Voice-2.1-Flash for low-latency agents at $15, 23 languages, one voice across all of them, in public preview on Azure.
Qwen-Audio 3.1
Alibaba Qwen's Sept 23, 2026 five-model audio stack — ASR, ASR-Next, TTS, TTS-Next and a full-duplex Realtime model — API-only on Model Studio, with list prices cut roughly 70% for TTS, 85% for Realtime and up to 95% for ASR.
ElevenLabs Music v2
The commercially safest AI music generator — built with label and publisher partnerships, with section-by-section editing, a full API and studio-grade exports.
Lyria 3
Google DeepMind's music generation model — high-fidelity 48kHz audio up to three minutes, available through Vertex AI and the Gemini API.
MiniMax Speech 2.8
MiniMax's text-to-speech family — expressive multilingual synthesis with paid-per-voice cloning and voice design, a large built-in voice library and async long-form jobs, at $60–$100 per 1M characters.
Stable Audio 2.5
Stability's music-and-sound model — fast, commercially-cleared instrumental tracks and sound design with an enterprise-friendly API.
Udio v4
The audio-quality leader among AI music tools — 48kHz rendering with the cleanest instrument separation, for film and brand-grade music.
ElevenLabs v3
ElevenLabs' previous-generation expressive text-to-speech model, superseded by Eleven v4 on September 28, 2026 — 70+ languages, multi-speaker dialogue and audio tags, now