Parakeet RNNT 1.1B
nvidia/parakeet-rnnt-1.1b
published Dec 2023 · updated Jun 2026
Parakeet RNNT 1.1B is an automatic speech recognition model that transcribes English speech into lower-case text using a FastConformer Transducer architecture with 1.1 billion parameters.
specs
| Task | Automatic Speech Recognition (ASR) |
| Architecture | FastConformer Transducer (RNNT) |
| Parameters | 1.1B |
| Language | English |
| License | CC-BY-4.0 |
about this model
Parakeet RNNT 1.1B is an automatic speech recognition (ASR) model that transcribes English speech into lower-case text using a FastConformer Transducer architecture with approximately 1.1 billion parameters.
Architecture and Performance
The model is built on the FastConformer architecture, which is 2.8x faster than the original Conformer and supports transcription of long-form speech up to 11 hours using limited context attention with a global token. The architecture was accepted at ASRU 2023.
It was trained on over 64,000 hours of English speech (40,000 hours from private data and 24,000 hours from public sources including LibriSpeech, Fisher, Switchboard, WSJ, VCTK, VoxPopuli, Europarl-ASR, MLS, Common Voice, and People’s Speech). The model uses a SentencePiece Unigram tokenizer (vocabulary size 1024).
Word Error Rate (WER) with greedy decoding on standard benchmarks:
| Dataset | WER (%) |
|---|---|
| LibriSpeech test-clean | 1.46 |
| GigaSpeech | 9.96 |
| SPGI Speech | 2.47 |
| TED-LIUM v3 | 3.11 |
| VoxPopuli | 3.92 |
| Common Voice | 5.79 |
| Earnings-22 | 14.11 |
| AMI | 17.10 |
Licensing
This model is released under the CC-BY-4.0 license.
best for
- ·Transcribing long-form English audio (up to 11 hours per clip)
- ·Real-time captioning and voice command recognition
- ·Batch transcription of meetings, podcasts, or call recordings
FAQ
FastConformer Transducer (RNNT) with 1.1B parameters, designed for efficient streaming and long-form inference.
The model accepts 16 kHz mono-channel WAV audio files as input.
It outputs transcribed text as a string in lower-case English without punctuation.
It is released under the CC-BY-4.0 license.
Call the OpenAI-compatible endpoint on gigarouter using an API key; no local installation required.
We're benchmarking and onboarding Parakeet RNNT 1.1B as a hosted, OpenAI-compatible API. Sign in for free credit and be ready when it lands, or tell us you want it and we'll prioritize it.