Open-source speech-to-text models
22 speech-to-text models from 13 publishers publish their weights, so you can run them yourself. Licences differ, so check commercial use on each model.
The cheapest published USD price is MiMo-V2.5-ASR from Xiaomi MiMo at $0.07 per hour; 3 of 22 have a comparable price.
Of these, 3 stream over WebSocket or WebRTC and 1 has an announced shutdown date.
Prices are normalized: USD per hour of audio, cheapest first. Tick models to compare them side by side, or refine these filters.
Speech-to-text and realtime $ per hour of audio, log scale 236 priced
other models matches your filtersAll models by price
| Compare | Model | Price ↑ |
|---|---|---|
mimo-v2.5-asrMiMo-V2.5-ASRXiaomi MiMo, STT |
$0.07 per hour | |
horizon-universalHorizon UniversalShunya Labs, STT |
$0.23 per hour | |
voxtral-mini-transcribe-realtime-2602Voxtral Mini Transcribe Realtime (26.02)Mistral AI, STT, ends 2027-04-30 |
$0.36 per hour | |
ai4bharat/indic-conformer-600m-multilingualindic-conformer-600m-multilingualAI4Bharat, STT |
– | |
cohere-transcribe-03-2026Cohere TranscribeCohere, STT |
– | |
cohere-transcribe-arabic-07-2026Cohere Transcribe ArabicCohere, STT |
– | |
FunAudioLLM/Fun-ASR-Nano-2512Fun-ASR-Nano-2512FunAudioLLM (Alibaba Tongyi), STT |
– | |
FunAudioLLM/SenseVoiceSmallSenseVoiceSmallFunAudioLLM (Alibaba Tongyi), STT |
– | |
chinda-asr-realtime-v1OpenThai ASR RealtimeiApp Technology, STT |
– | |
microsoft/Phi-4-multimodal-instructPhi-4-multimodal-instructMicrosoft (open weights), STT |
– | |
microsoft/VibeVoice-ASRVibeVoice-ASRMicrosoft (open weights), STT |
– | |
microsoft/VibeVoice-ASR-BitNetVibeVoice-ASR-BitNetMicrosoft (open weights), STT |
– | |
microsoft/VibeVoice-ASR-Streaming-7BVibeVoice-ASR-Streaming-7BMicrosoft (open weights), STT |
– | |
nvidia/nemotron-asr-streamingNemotron ASR StreamingNVIDIA, STT |
– | |
nvidia/nemotron-transcribeNemotron 3.5 TranscribeNVIDIA, STT |
– | |
nvidia/parakeet-tdt-0_6bParakeet TDT 0.6B v3NVIDIA, STT |
– | |
nvidia/parakeet-tdt-0_6b-v2Parakeet TDT 0.6B v2NVIDIA, STT |
– | |
rumble-2-flashRumble-2-FlashReson8, STT |
€0.0114 per hour | |
ai-sage/GigaAM-MultilingualGigaAM-MultilingualSber SaluteSpeech, STT |
– | |
ai-sage/GigaAM-v3GigaAM-v3Sber SaluteSpeech, STT |
– | |
Sunbird/faster-whisper-51-african-languagesSunbird ASR (Whisper large-v3, 51 African languages)Sunbird AI, STT |
– | |
t-tech/T-oneT-oneT-Bank VoiceKit, STT |
– |
Speech-to-text by feature
- Open-source / open weights
- Custom vocabulary
- Language detection
- PII redaction
- Profanity filter
- Punctuation
- Sentiment
- Speaker diarization
- Summarization
- Topic detection
- Translation
- Word timestamps
Speech-to-text by language
- Afrikaans
- Albanian
- Amharic
- Arabic
- Armenian
- Assamese
- Azerbaijani
- Basque
- Belarusian
- Bengali
- Bosnian
- Bulgarian
- Burmese
- Cantonese
- Catalan
- Chinese
- Croatian
- Czech
- Danish
- Dutch
- English
- Estonian
- Filipino
- Finnish
- French
- Galician
- Georgian
- German
- Greek
- Gujarati
- Hausa
- Hebrew
- Hindi
- Hungarian
- Icelandic
- Indonesian
- Irish
- Italian
- Japanese
- Javanese
- Kannada
- Kazakh
- Khmer
- Korean
- Lao
- Latvian
- Lithuanian
- Macedonian
- Malay
- Malayalam
- Maltese
- Marathi
- Mongolian
- Nepali
- Norwegian
- Odia
- Pashto
- Persian
- Polish
- Portuguese
- Punjabi
- Romanian
- Russian
- Serbian
- Sinhala
- Slovak
- Slovenian
- Somali
- Spanish
- Swahili
- Swedish
- Tagalog
- Tamil
- Telugu
- Thai
- Turkish
- Ukrainian
- Urdu
- Uzbek
- Vietnamese
- Welsh
- Yoruba
- Zulu
- ak
- ast
- ba
- bgc
- bhb
- bho
- bo
- br
- brx
- ceb
- cjy
- ckb
- doi
- eo
- ff
- fo
- fy
- gan
- gom
- hak
- haw
- hne
- hsn
- ht
- ia
- ig
- iw
- kab
- kam
- kea
- kok
- ks
- ky
- la
- lb
- lg
- ln
- luo
- mai
- mg
- mhr
- mi
- mni
- nan
- nb
- nn
- nr
- nso
- ny
- oc
- om
- rup
- rw
- sa
- sat
- sd
- sn
- ss
- st
- su
- tg
- tk
- tn
- ts
- tt
- ug
- ve
- wo
- wuu
- xh
- yi