OpenBMB: VoxCPM2
openbmb/voxcpm2Hugging FaceSpeechardiffusionenmultilingualmytext-to-speechttsvoice-cloningvoice-designvoxcpmzh
Reference inputs & limits
Accepted uploads and guidance from each provider. Limits apply to one generation.
ImpCC · speech generation · up to 1 file · 5 MiB total
audio as reference (up to 1) · audio/wav, audio/mpeg, audio/flac, audio/ogg, audio/mp4, audio/webm
One audio sample chooses a custom voice instead of the named voice. Samples must total at most 5 MiB. A sample takes precedence over voice and does not create a saved speaker profile.
Providers
Live endpoints for this model. Response is the time from request to result; uptime is measured over the last day and week. Nothing here streams, so there is no first token to time and no token rate to report.
| Provider | Per M chars | Response | Capacity | Uptime 1d | Uptime 7d | Health | Data |
|---|---|---|---|---|---|---|---|
| ImpCCbf1690% offCA | $1.50 | 555ms30m | 4/4 free | 100.0% | 97.7% | Healthy | ZDR |
Pricing
USD per million characters synthesized.
| Provider | Per M chars |
|---|---|
| ImpCCbf1690% off | $1.50 |
Price history
Price · USD per million characters
Performance
Percentiles cover the labelled window: the last 30 minutes when it has enough samples, otherwise the last day. Response is the time in seconds from request to result. This model answers whole rather than streaming, so there is no first token to time and no rate between tokens to report — response time is the whole measurement there is.
| Provider | Response p50 | p75 | p90 | p99 |
|---|---|---|---|---|
| ImpCCbf16 | 555ms30m | 625ms | 641ms | 676ms |
Uptime
Share of successful requests and probes over the last day and week.
- ImpCCbf161d 100.0%7d 97.7%Healthy
Availability by hour · last 7 days
Share of successful requests and probes per hour and provider. The line spans hours with nothing to measure; where samples are sparse, a dot marks each hour actually measured.
Activity
Speech requests per UTC day over the last 30 days. Every billed request counts, the platform's own paid probe traffic included.
curl https://api-staging.heimdallm.com/api/v1/audio/speech \ -H "Authorization: Bearer $HEIMDALL_API_KEY" \ -F 'model=openbmb/voxcpm2' \ -F 'input=Hello, this is a new sentence in the reference voice.' \ -F 'file=@voice.wav;type=audio/wav' \ -F response_format=wav \ --output speech.wavSet HEIMDALL_API_KEY to a key from your account. See the quickstart for routing suffixes and provider preferences.
About
VoxCPM2 is a tokenizer-free, diffusion autoregressive Text-to-Speech model — 2B parameters, 30 languages, 48kHz audio output, trained on over 2 million hours of multilingual speech data.
- Author
- OpenBMB
- Family
- openbmb/voxcpm2
- Weights
- openbmb/VoxCPM2
- Parameters
- input_references (list), language (30 choices), max_new_tokens (3–4096), response_format (wav, mp3, opus, flac, pcm), speed (0.25–4), voice (default)
- Licence
- apache-2.0
- Size
- 2.3B parameters, BF16 weights
- Tokenizer
- —
- Available until
- —
- Voices
- default
From the model card
What the repository states about these weights, read from openbmb/VoxCPM2 on Oct 11, 2026.
- Languages
- Chinese, English, Arabic, Burmese, Danish, Dutch, Finnish, French, and 22 more
- Papers
- 2509.24650
- Datasets
- —
- Revision
- 32279efAug 18, 2026
- Downloads
- 633.7K
- Likes
- 1.7K