I've migrated the ChatVibes Text-to-Speech engine to Speech 2.6 Turbo (MiniMax via 302.ai). This model succeeds Speech-02 and provides improvements in latency, prosody, and multilingual handling.
What's New
Audio Quality and Prosody
Speech 2.6 produces smoother phrasing and more stable prosody than Speech-02. It also handles structured text (numbers, URLs, dates) more cleanly.
Language Coverage
Speech 2.6 supports a broad multilingual set (40+ languages and dialects, per MiniMax documentation). Performance varies by language, so users who rely on non-English output may want to test their specific cases.
Voice Compatibility
Not all existing voice IDs are supported by Speech 2.6. Supported voices have already been upgraded automatically. Unsupported voices continue using the legacy Speech-02 engine.
Reliability and Fallback
If Speech 2.6 encounters service issues, the system automatically falls back to Speech-02 (via Wavespeed) to maintain continuity.
What You Need To Do
No action is required.
- Your current voice settings carry over.
- Supported voices now use Speech 2.6.
- Unsupported voices remain on Speech-02.
- Language options are available in Viewer Preferences and the Channel Dashboard.