Current Soniox text-to-speech model, generally available since August 11, 2026, adding audio tags for emotion and delivery, voice cloning and in-sentence language switching across 60+ languages; it is API-compatible with v1.
$11.985 / 1M characters
Current Soniox text-to-speech model, generally available since August 11, 2026, adding audio tags for emotion and delivery, voice cloning and in-sentence language switching across 60+ languages; it is API-compatible with v1.
Live-audio flagship from Soniox, released June 16, 2026, streaming transcription in 60+ languages with rebuilt speaker separation, faster semantic endpointing for voice agents and live translation across 3,600+ language pairs.
Soniox's current async model for recorded audio, released June 11, 2026, covering 60+ languages with reengineered speaker separation, better language identification and stronger context biasing for names, product terms and custom vocabulary.
Soniox's first generally available TTS model (April 29, 2026), streaming speech in 60+ languages with accurate reading of phone numbers, emails and IDs. Deprecated by tts-rt-v2 and removed August 31, 2026, after which requests route to v2.
Preview of Soniox's first text-to-speech model, released April 23, 2026, streaming audio in 60+ languages before a sentence finishes. Since April 29, 2026 the id is an alias for tts-rt-v1, which now routes to tts-rt-v2.
Soniox v4 Real-Time, released February 5, 2026, which cut endpoint and finalization latency, improved semantic endpointing for turn-taking and added max_endpoint_delay_ms. Replaced by stt-rt-v5; since June 30, 2026 requests to this id are served by v5.
Soniox v4 Async, released January 29, 2026, which Soniox said reached human-parity transcription with better diarization and multi-hour handling. Replaced by stt-async-v5; since June 30, 2026 this id is an alias served by v5.
Soniox's v3 async model, released October 21, 2025, which brought translation to file transcription along with structured context and audio up to 5 hours in 60+ languages. Superseded by stt-async-v4 and removed on February 28, 2026.
Soniox v3 real-time model, released October 21, 2025, streaming transcription and translation in 60+ languages with better mid-sentence language switching, overlapping-speech diarization and 5-hour sessions. Replaced by stt-rt-v4 and removed February 28, 2026.
Second real-time preview from Soniox, released August 5, 2025 with higher accuracy, translation across all language pairs and more reliable automatic language switching. Retired October 31, 2025, when the name was aliased to stt-rt-v3.
Original alias for Soniox's async file-transcription preview, pointing to stt-async-preview-v1. That preview was retired on October 31, 2025 and its name aliased to the v3 models; stt-async-v5 is the current async model.
First-generation async model from Soniox's 2025 preview API, transcribing uploaded files with speaker labels in 60+ languages. Replaced by stt-async-v3 on October 21, 2025 and retired ten days later, when its name became an alias for v3.
Original name of Soniox's real-time preview, later an alias that tracked the newest real-time preview (stt-rt-preview-v2 by September 2025). The preview line was retired on October 31, 2025; stt-rt-v5 is the current real-time model.
Floating alias that Soniox documented in late 2025 as always pointing to its latest active real-time model, then stt-rt-v3. It no longer appears in Soniox's alias list; stt-rt-v5 is the current real-time model.