LocalClaw/TTS/Nemotron 3.5 Transcribe 30B A4B
Local audio-language model

Nemotron 3.5 Transcribe 30B A4B

Catalogue summary: Offline speech and audio model for transcription in 40 languages, speech-to-text translation, speaker-attributed transcripts, word timestamps and audio questions. NVIDIA publishes complete weights and local Transformers, NeMo Speech/vLLM and SGLang-Omni inference paths. Supports recordings up to one hour; outputs text only.

Repository editorial metadata; verify comparative claims in the linked upstream material.

GPU recommendedspeech-to-text transcription40 languagesOpenMDW-1.1 (model weights); NeMo Speech code Apache 2.0
Choose an app

Start with the recommended app, or choose another supported option.

Compare speech models
Open model filesModel weights and official local setupOpen in UnslothLaunches Unsloth Desktop on this modelOpen Python packageVerified package page without exposing a commandOpen official projectApplication source and guided setupOpen on Hugging FaceFiles, licence and model card

Desktop app links require the app to be installed. If nothing opens, LocalClaw will show app-download and model-file fallbacks.

Catalogue quality
Not rated
Catalogue speed
Not rated
Model size
34B total / 4.3B active including audio encoder; 67.496 GB reference safetensors
Voices
N/A: text output only; no voice cloning, speech synthesis or real-time streaming ASR

Can Nemotron 3.5 Transcribe 30B A4B run locally?

Nemotron 3.5 Transcribe 30B A4B can run locally for offline speech-to-text. Use the verified setup options on this page; no terminal command is required to choose the right path.

OpenMDW-1.1 (model weights); NeMo Speech code Apache 2.0 license. Still verify upstream usage notes before shipping.

Upstream source
multilingualspeech-translationaudio-languagelong-formdiarization

Model details

Type
Local audio-language model
Family
nemotron
Latency
not rated
Formats
pytorchtransformerssafetensorsvllm
Languages
cs, nl, en, fr, de, hi, it, ko, pl, pt, ru, sk, es, bg, hr, da, et, fi, hu, ja, lv, lt, ro, sl, sv, th, tr, uk, ar, bn, zh, el, he, id, mt, mr, tl, ta, te, vi
Context
Official HF release dated October 28, 2026; runtime source: github.com/NVIDIA-NeMo/Speech. Reference checkpoint is 67,496,013,734 bytes. The documented H100 80 GB inference configuration is a supported deployment, not a measured universal minimum or a system-RAM claim. FP8 and NVFP4 are official variants of this same model; hardware support varies by precision (L40S 48 GB FP8; RTX 5090 NVFP4). Do not apply those configurations to BF16. Language/task coverage varies: consult the official language matrix for translation targets. Word timestamps use the NeMo/vLLM recipe; SGLang-Omni currently covers whole-file transcription only, without translation, speaker-attributed output or timestamps. Internal chunking is not streaming ASR. NIM optimized throughput is not the open-source runtime throughput. No local inference test performed; quality and speed remain unrated.

Install locally

01
Check runtimeConfirm the backend supports pytorch, transformers, safetensors, vllm on your machine.
02
Open recommended setupUse the app and model links above. LocalClaw does not expose a terminal command.
03
Test locallyRun a short private audio prompt before moving into production workflows.

Good for

  • speech-to-text transcription
  • GPU recommended local workflows
  • multilingual, speech-translation, audio-language

Watch before shipping

  • Validate pronunciation, latency and artifacts with your own voice samples.
  • Review the upstream license and acceptable-use notes.
  • Benchmark on your target CPU, Apple Silicon or GPU setup.

Related TTS and speech models

NVIDIA Nemotron 3.5 ASR Streaming 0.6B Local ASR model · Q 9.4 · Speed 9.3 IBM Granite Team Granite Speech 4.1 2B Plus Local ASR model · Q 9.3 · Speed 7.8 OpenMOSS / MOSI.AI MOSS-Transcribe-Diarize 0.9B Local ASR model · Q 9.4 · Speed 8.4 OKESTRO AI Lab / Lee Junseok FastSLM-ASR Local ASR model · Q 9.2 · Speed 7.8 OpenMOSS / MOSI.AI MOSS-Audio 4B Instruct Local ASR model · Q 9.2 · Speed 7.7 FireRed Team FireRedAudio Local ASR model · Q 9.5 · Speed 6.6 ai-sage GigaChat Audio 10B A1.8B Local ASR model · Q 9.1 · Speed 6.7 NVIDIA Parakeet TDT 0.6B v3 Local ASR model · Q 9.5 · Speed 9.7
CompareBrowse all TTS models Local AIBrowse LLM models macOS appGet LocalClaw