MiniMax Speech 2.8 HD
Speech 2.8 HD is MiniMax's text-to-speech model, announced on 23 January 2026 in HD and Turbo variants, and the one that topped the blind speech arenas at launch ahead of OpenAI and ElevenLabs. It takes an emotion setting, pause markers written into the text, and 32 languages, with cloning on MiniMax's own platform. On fal it costs $0.10 per thousand characters with a handful of named stock voices.
Anvisha Pai, Co-founder & CEO, Voyager
MiniMax · released
›What we asked for
Hi, I'm MiniMax Speech 2.8 HD, a voice model from MiniMax in Shanghai. I speak thirty-two languages and take an emotion setting, but today I'm reading plain text. Here is the same script every voice on this site reads, so you can compare us.
fal-ai/minimax/speech-2.8-hd
output_format: url, voice_id: Wise_Woman, speed: 1
returned 18.0 seconds of mp3
At a glance
- Price
- $0.10 per thousand characters on fal, the same as Eleven v3; our four scripts cost about ten cents.
- Voices
- Named stock voices on fal such as Wise_Woman, which we used; cloning on MiniMax's platform.
- Emotion
- A setting with seven values from happy to neutral; our scripts left it unset.
- Languages
- 32.
- Pauses
- Write <#1.5#> in the text for a 1.5-second pause.
Anvisha's take
We have not scored the four readings by ear; press play and judge the voice yourself. What we measured: MiniMax Speech 2.8 HD read the four scripts in 12 to 25 seconds of audio each and returned in four to six seconds, about twice as fast as Eleven v3 at the same $0.10 per thousand characters. The Wise_Woman stock voice was used with the emotion setting left off, so the emotional script is read from the words alone. It defaults to returning audio as hex; we set output_format to url. This is the model that topped the blind speech arenas at launch, so the listening test is the one that matters.
MiniMax Speech 2.8 HD real output: the same 4 scripts every model gets
Every voice model on this site gets the same 4 scripts, so you can compare like with like. These are the readings MiniMax Speech 2.8 HD returned, one try each, nothing re-rolled or retouched. Generated on . The 4 readings took 4 to 6 seconds each and cost $0.0976 in total.
Pacing, warmth and whether the pauses land where a voice actor would put them.
›What we asked for
Meet the Ember mug. It keeps your coffee at exactly the temperature you choose, for up to ninety minutes, and it tells your phone when it's ready. Warm from the first sip to the last. Ember. Coffee, on your terms.
fal-ai/minimax/speech-2.8-hd
output_format: url, voice_id: Wise_Woman, speed: 1
returned 15.6 seconds of mp3
Steady narration over a long paragraph, clean list rhythm and a natural close.
›What we asked for
A pour-over works in four steps. First, rinse the paper filter with hot water so it doesn't taste of paper. Second, add the grounds and pour just enough water to wet them, then wait thirty seconds while they bloom. Third, pour the rest of the water in slow circles. Fourth, wait about three minutes for it to drip through. That's it: a cup that tastes like the coffee, not the machine.
fal-ai/minimax/speech-2.8-hd
output_format: url, voice_id: Wise_Woman, speed: 1
returned 25.2 seconds of mp3
Range: excitement, then a hushed aside, from the words alone with no tags.
›What we asked for
We did it! We actually did it! I can't believe it's finally over. Okay. Okay, deep breath. Let's not tell anyone until Monday, alright? Just... let me have this one quiet night.
fal-ai/minimax/speech-2.8-hd
output_format: url, voice_id: Wise_Woman, speed: 1
returned 12.1 seconds of mp3
Reading numbers, times, currency, an email address and two hard names correctly.
›What we asked for
Your order ships on the 24th of September 2026 and arrives by 9:15 a.m. The total is $1,249.99, including VAT. Questions? Ask for Siobhan Nguyen or Ravi Parikh on 0800 555 0199, or email help@moda.app.
fal-ai/minimax/speech-2.8-hd
output_format: url, voice_id: Wise_Woman, speed: 1
returned 24.7 seconds of mp3
How the cost under each picture was worked out: fal's MiniMax Speech 2.8 HD rate of $0.10 per 1,000 characters (fal model page, read 2026-09-10) for the 213 characters sent; fal does not return the charge.
MiniMax Speech 2.8 HD pricing
| Where | Unit | Price | Source |
|---|---|---|---|
| fal | 1,000 characters | $0.10 | fal model page, read |
Is MiniMax Speech 2.8 HD free?
MiniMax's own platform and the Hailuo Audio app give new accounts trial credits; the API charges per character. fal charges per character with no free tier.
- •MiniMax platform: trial credits for new accounts.
- •fal: per character.
MiniMax platform, read
Where to use MiniMax Speech 2.8 HD
- MiniMax platform · MiniMax's API with cloning and both HD and Turbo variants.
- fal · $0.10 per thousand characters, stock voices.
- Voyager · Runs inside Voyager, the agent for creative work. Private preview by waitlist.
Voyager
Run MiniMax Speech 2.8 HD inside Voyager
Voyager is an agent for creative work that runs every model in this reference, MiniMax Speech 2.8 HD included. Brief it, and it picks the model, makes the thing and hands back something you can edit. Create an account to get started.
What we noticed running MiniMax Speech 2.8 HD
- •25.2 seconds for the paragraph, the second longest read of the four after Gemini at 25.8, returned in five seconds. See the explainer paragraph reading.
- •The endpoint returns hex-encoded audio unless output_format is set to url; the harness sets it.
- •The emotion parameter was left unset on every script so the comparison stays on the same words.
MiniMax Speech 2.8 HD limits and API parameters
| Limit | Value |
|---|---|
| Voices on fal | Stock voices only; cloning is on MiniMax's platform. fal API reference, read |
| Output | Defaults to hex-encoded audio; set output_format to url. fal API reference, read |
›Every setting, for developers
fal (fal-ai/minimax/speech-2.8-hd)
fal API reference, read
| Parameter | Type | Values | Default |
|---|---|---|---|
| prompt | string | The script, with optional <#n#> pauses Required. | |
| voice_id | string | Named stock voices | Wise_Woman |
| speed | number | 0.5 to 2 | 1 |
| emotion | enum | happy, sad, angry, fearful, disgusted, surprised, neutral | |
| vol | number | 0 to 10 | 1 |
| pitch | integer | -12 to 12 | |
| sample_rate | enum | 8000 to 44100 | 32000 |
| format | enum | mp3, pcm, flac | mp3 |
| output_format | enum | url, hex | hex |
What MiniMax Speech 2.8 HD will not say
MiniMax's platform terms bar cloning a voice without the speaker's consent, impersonation, fraud and sexual content involving minors, and the hosted endpoints run under fal's partner terms for commercial use.
MiniMax platform, read
Alternatives to MiniMax Speech 2.8 HD
Frequently asked questions
What is MiniMax Speech 2.8?
MiniMax's text-to-speech model from 23 January 2026, in HD and Turbo variants, with 32 languages, an emotion setting and voice cloning. It led the blind speech arenas at launch.
How much does MiniMax Speech 2.8 HD cost?
$0.10 per thousand characters on fal, read on 10 September 2026.
How this page is made
Prices, limits and settings are read from the linked vendor and host pages on the dates shown. Every picture was generated by us through fal with the request shown under it, and the original files are kept unedited. A generation that failed is shown as a failure. Model and vendor names are used to identify the products; no vendor imagery appears on this page.
