MiniMax · Audio model
Highest-fidelity MiniMax voices, emotion control
Compare credit costs from the published model catalog. The studio shows the current quote for your settings before you generate.
| Text length | Credits |
|---|---|
| Minimum charge | 0.5 |
| Per 1,000 characters | 1 |
| Example — a 5,000-character script | 5 |
Use this text-to-speech model for narration, explainers or a product walkthrough. Enter the words to speak, up to 5,000 characters per request. No source photo, recording or cloned voice is required. Start with a paragraph containing your difficult names and numbers before producing a longer script.
Edit the script and punctuation to guide delivery. The provider documents pause tags such as <#0.5#> and interjections such as (laughs); audition these in a short passage. King AI currently uses the provider's default voice settings for this model. The studio's general vocal-type selector does not select a MiniMax voice; named voices, pitch, speed and emotion sliders from the provider API are not exposed here.
HD and Turbo are separate speech variants, not video models. Compare a short passage before deciding whether HD suits your narration. HD costs 0.5 credits for 250 characters, 1 for 1,000 and 5 for 5,000; Turbo costs 0.6 for 1,000. Pricing is proportional to submitted characters with a minimum charge, not rounded up in 1,000-character blocks. Each new generation has its own quote.
Listen for names, pauses and unwanted spoken instructions. Rewrite a mispronounced word phonetically or split the script into shorter sections, then check continuity in your audio editor. If a request fails, read its status and credit balance before retrying; a new submission is a separate request. A voiceover is an audio asset, so assemble it with footage and captions in your editing workflow.
Provider reference: MiniMax Speech 2.8 HD API. Available King AI controls and credit costs are described above.
Compare MiniMax Speech 2.8 Turbo pricing