5 AI speech and audio models from 4 labs. 2 have weights you can download. Input prices run from $0.50 to $2 per 1M tokens. The largest context window is 96K tokens (Gemini 3.5 Transcribe). Newest check: 2 Nov 2026.
94 of 197 models were checked against the vendor's own page in the last 30 days.
| Model | Lab | Context | In / out per 1M | Reads and writes | Status | Checked |
|---|---|---|---|---|---|---|
| Gemini 3.8 Flash TTSGemini 3.8 Flash TTS (Google DeepMind, September 2026) turns text into expressive one- or two-speaker audio, with voice design and voice replication built in. | 8K | $0.50 / $9 | text → audio | Available | 11 Oct | |
| Suno V6Suno's flagship music model, launched September 9, 2026 and trained on catalogs licensed from Warner Music, BMG and Believe. | Suno | -- | -- | text, audio, image, video → audio | Available | 13 Sep |
| Gemini 3.5 TranscribeGemini 3.5 Transcribe (Google DeepMind, Aug 2026) is a dedicated speech-to-text model with built-in diarization, timestamps, and disfluency cleanup across dozens of languages. | 96K | $2 / $12 | audio → text | Preview | 27 Aug | |
| MiniMax Music 3.0MiniMax Music 3.0 is MiniMax's open-weight text-to-music model, released in 2026, generating complete 5-minute songs with performed, not synthesized, vocals. | MiniMax | -- | -- | text → audio | Available | 25 Sep |
| Nemotron 3.5 ASRNVIDIA's open streaming speech recognizer, released 4 June 2026, transcribing dozens of language-locales from a single checkpoint. | NVIDIA | -- | -- | audio → text | Available | 2 Nov |
A person checks every record against the vendor's own page. Rankings cannot be bought. How Hokai checks records. Not sure what you need? Describe the job to Smart Match and get three picks with reasons.