MiniMax Speech 2.8 HD when the read has to breathe where you say it does
This is the one engine here where the mood is a setting rather than a slider to guess at or a tag to remember: you pick "happy" or "neutral" from a list of seven and that is the whole of it. The other thing nobody else has is a pause you write into the script to the hundredth of a second, so the silence before the offer lands where you want it rather than where the engine saw a full stop.
Rendered by this model
Turn the sound on and listen
The script the engine was given
Nowa kolekcja jest już w sklepie. <#0.7#> Bawełna, len i jeden kolor, który schodzi pierwszy. <#0.5#> Wejdź, obejrzyj i zamów — wysyłamy w dwadzieścia cztery godziny.
Specification
- Built by
- MiniMax
- Speaks Polish
- yes — checked on a Polish sentence
- Ready-made voices
- 17
- Steering the delivery
- an emotion picked from a list, as a setting
- Character limit
- 10,000
- File from the engine
- mp3
- Render time
- about 10s
- Cost
- 1 credit per 1000 characters, rounded up
What it is good for
- The mood is a setting, not a guess
- Seven states to choose from: happy, sad, angry, fearful, disgusted, surprised and neutral. You do not describe the mood in the script or hunt for the right slider value — the same advert recorded happy and recorded neutral gives you two takes to compare without changing a word.
- Pauses exactly where you put them
- A marker in the text holds the read for a stated number of seconds, from a hundredth up to ninety-nine. It is the only way to make a read fit a cut you have already made, rather than cutting to whatever the engine did with your punctuation.
- Ten thousand characters in one job
- The highest limit of the MiniMax pair and one of the highest anywhere on this list. A full product description goes in as a single read, so there is nothing to stitch and the voice does not change halfway through.
When to pick something else
Abbreviations and shorthand are read its own way, and that is the one thing to watch here. The same family read "330 ml" back to us as "330 milinów" and "do 14:00" as "do 14 zero", which is why the number in the sample script is spelled out. If your copy is full of clock times, units and dimensions and you would rather not expand them by hand, ElevenLabs Multilingual v2 does it for you — it turned "24 h" into "24 godziny" correctly. If all you want is a shorter wait, MiniMax Speech 2.8 Turbo does the same job at the same price per character.
See every modelHow to brief this model
The things this engine handles differently from the rest. The habits that hold for every model are collected separately, in the help centre.
- Write the pause in as a marker, for instance after the opening line. The engine performs it as silence — we checked that it is not read aloud — and it is the only way to make a read land inside a cut you have already timed.
- Spell out clock times, units and dimensions. This engine gave us "330 milinów" for "330 ml" and "14 zero" for "14:00"; "trzysta trzydzieści mililitrów" comes out right.
- Pick the emotion from the field and do not describe it in the script. A mood written into a sentence gets read as part of the advert, because the field next to it already does that job.
- Leave the language field set to Polish. Without it the engine reads Polish text with an English mouth, which is the same difference the four English-only pages on this list are about.
- Use this tier rather than Turbo for anything you publish. The credit is the same and the difference is in the recording itself — Turbo is for auditioning script variants.
Questions about this model
Other models
- MiniMax Speech 2.8 TurboThe same Polish voice as HD, back sooner. For auditioning takes.
- ElevenLabs Multilingual v2It expands "24 h" into "24 godziny" by itself. For copy with numbers in it.
- ElevenLabs Turbo v2.5The shortest wait of the ten. For short adverts and quick takes.
- ElevenLabs v3You write the mood into the middle of the sentence, in square brackets.
- VibeVoice 7BTwo people talking in one file. In Polish.
- KokoroEnglish only, and nothing to configure. For adverts abroad.
- Chatterbox HDThe only engine here that will speak in your own voice.
- Orpheus TTSA laugh or a sigh mid-sentence. English, and slowly.
- DiaTwo people out of one text field. No voice to choose.
Try this model on your own product
Sign up without a card and get 40 credits. Switching model is one click, so you can send the same brief through several of them and keep the best frame.