Skip to content
Model infoModelsSnapshot: 15 Oct 2026

Xiaomi / MiMo / MiMo speech synthesis

MiMo V2.5 TTS

Speech synthesis with built-in voices, singing and natural-language control of expression.

Speech & voice
availableClosed weightsSpeech generation
Official documentation (opens in a new tab)Reviewed 14 Oct 2026
16 of 24 specification fields collected
Output formatsWAV, MP3, PCM16See specification notes below
Languages2 listedFull language list in specifications
Input billingSee billing notesXiaomi MiMo API
Output billingSee billing notesAccess and billing notes below

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
8KSource for Context / input limit (opens in a new tab)

Shared table cells span all three TTS variants.

Maximum output
8KSource for Maximum output (opens in a new tab)

Provider token limit, not a conversion into audio duration.

Input modalities
TextSource for Input modalities (opens in a new tab)
Output modalities
AudioSource for Output modalities (opens in a new tab)
Reasoning
Not supportedSource for Reasoning (opens in a new tab)

TTS API states that reasoning_tokens is always zero.

Tool calling
Not collected
Structured output
Not collected
Streaming
SupportedSource for Streaming (opens in a new tab)

Incremental low-latency streaming has been restored.

Batch processing
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
2026-04-23Source for Release date (opens in a new tab)
Parameters
Not collected
Architecture
Not collected
Languages
Chinese, EnglishSource for Languages (opens in a new tab)

Documented voices and bilingual voice descriptions; wider language/accent capability is not an exhaustive supported-language guarantee.

Context caching
Not collected

TTS response schema includes cached_tokens, but a model-specific cache policy or discount was not established. Do not inherit language-model tariffs.

Output formats
WAV, MP3, PCM16Source for Output formats (opens in a new tab)

pcm and pcm16 select PCM16. Streaming output should use PCM16.

outputSampleRate
24 kHz PCM16LE mono for the streaming examplesSource for outputSampleRate (opens in a new tab)

Do not infer the same container, channel layout or sample-rate setting for every non-streaming format.

singing
SupportedSource for singing (opens in a new tab)
Voice cloning
Not supportedSource for Voice cloning (opens in a new tab)
Voice design
Not supportedSource for Voice design (opens in a new tab)
builtInVoices
SupportedSource for builtInVoices (opens in a new tab)
Style controls
Natural-language directions and inline tags for delivery, emotion, rhythm and vocal eventsSource for Style controls (opens in a new tab)
Published price
Free for a limited timeSource for Published price (opens in a new tab)

No end date or post-promotion rate verified. Not a permanent zero-cost tariff.

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Xiaomi MiMo API

Provider source (opens in a new tab)
API model ID
mimo-v2.5-tts
Access
API key required. Current temporary free offer; rate limits and model-specific streaming restrictions apply.
Regions
API cluster/account specific; no residency or jurisdiction guarantee inferred.
API endpoint URL
POST https://api.xiaomimimo.com/v1/chat/completionsSource for API endpoint URL (opens in a new tab)
Authentication
api-key header or Bearer authenticationSource for Authentication (opens in a new tab)
rateLimit
100 requests/minute; 10M tokens/minuteSource for rateLimit (opens in a new tab)

Published per-model limits; not a latency or throughput guarantee.

Pricing terms
Temporary free access; subsequent pricing and end date unverifiedSource for Pricing terms (opens in a new tab)
streamingTransport
SSE with incremental audio chunksSource for streamingTransport (opens in a new tab)

Endpoint pricing has not been collected. Check the provider for applicable rates.

Data & enterprise terms

Retention, training use, regional processing and enterprise terms have not been summarized for this endpoint.

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

Xiaomi MiMo API

mimo-v2.5-tts
Not verified
How it works
Not collected
Minimum reusable input
Not collected
Lifetime & refresh
Not collected
How to enable it
Not collected
Usage fields
usage.prompt_tokens_details.cached_tokens exists in the TTS response schemaSource for Usage fields (opens in a new tab)

A schema field does not establish an applicable cache discount or lifetime.

Billing rules
Not collected
Limits & data boundaries
Not collected

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Voice choices

9 voices

Choose a voice within MiMo V2.5 TTS. These are API voice identifiers, not additional model releases. Source links document each voice.

Published voice catalog
Voice / API IDLanguage & accentDelivery
MiMo defaultmimo_defaultCluster-dependentChina uses 冰糖; other clusters use Mia. This selector is not a ninth distinct voice.Voice source (opens in a new tab)
冰糖冰糖ChineseFemaleVoice source (opens in a new tab)
茉莉茉莉ChineseFemaleVoice source (opens in a new tab)
苏打苏打ChineseMaleVoice source (opens in a new tab)
白桦白桦ChineseMaleVoice source (opens in a new tab)
MiaMiaEnglishFemaleVoice source (opens in a new tab)
ChloeChloeEnglishFemaleVoice source (opens in a new tab)
MiloMiloEnglishMaleVoice source (opens in a new tab)

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Inferred closed-weight classification for this exact hosted MiMo release from the original release description and developer public-weight inventory. The matching weights were not found. Later MiMo V2.5/V2.6 language weights and MiMo V2.5 ASR are explicitly public and are not substituted for this release; public MiMo-Audio components do not establish public weights for the separate TTS service.

Classification evidence (opens in a new tab)

Lifecycle dates

Series released23 Apr 2026Source (opens in a new tab)

Sources reviewed