# NVIDIA Nemotron Transcribe handles meetings

> An offline model adds speaker-attributed transcription.

By BIGCHANGE.AI Editorial

Published: 2026-10-29T13:22:39.265Z
Updated: 2026-10-29T13:22:39.265Z
Canonical: https://bigchange.ai/blog/nvidia-nemotron-3-5-transcribe-speaker-prompts

![Conceptual charcoal illustration of three people around a meeting table, with two speaking and a small unbranded audio recorder between them.](https://bigchange.ai/api/media/file/nemotron-conversation-hero-v2.png)
AI-generated conceptual illustration by BIG CHANGE.

NVIDIA's [Nemotron 3.5 Transcribe](https://research.nvidia.com/labs/nemotron/NVIDIA-Nemotron-3.5-Transcribe/), released October 28, transcribes recorded audio up to one hour in 40 languages. It accepts instructions and can label speakers and return word timestamps.

## The big change

- A transcript can serve different purposes: a verbatim record preserves hesitations, while a readable version can omit them. A team needs to choose the form its readers require.
- Speaker labels help people follow a conversation, but do not establish a person's real-world identity. Check attributed speech against the recording and names against reliable records.
- A benchmark compares output with a defined reference. Its score is useful evidence for that test, while the accuracy of a particular meeting depends on the audio and task.

## From recording to transcript

NVIDIA combines a speech encoder and frozen Nemotron 3 Diarization component with the Nemotron 3.5 Lightning language model. Its [model card](https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B) describes prompts for style, vocabulary and speaker tags; timestamps are aligned after generation. Context terms are hints, not evidence of identity.

The model uses the [OpenMDW 1.1 license](https://openmdw.ai/license/1-1/). NVIDIA documents an H100 80 GB BF16 profile, L40S 48 GB FP8 and RTX 5090 NVFP4. BIG CHANGE reviewed the documentation and did not run the model.

## Reading the results

NVIDIA reports 4.71% mean word error on nine English leaderboard sets; the next listed open-weight model scored 4.94%. These vendor-reported results are test-specific.

The [Open ASR Leaderboard methodology](https://huggingface.co/spaces/hf-audio/open_asr_leaderboard/blame/main/constants.py) measures substitutions, insertions and deletions against a reference and normalizes text before scoring. That score does not by itself establish whether speaker labels or a requested transcript style are correct.

BIG CHANGE previously covered the [standalone diarization model](https://bigchange.ai/blog/nvidia-nemotron-3-speaker-diarization-model). This is a separate release.

## Sources

- [NVIDIA Transcribe release](https://research.nvidia.com/labs/nemotron/NVIDIA-Nemotron-3.5-Transcribe/) — Architecture and results.
- [NVIDIA Transcribe model card](https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B) — Instructions, deployment and license.
- [Open ASR Leaderboard methodology](https://huggingface.co/spaces/hf-audio/open_asr_leaderboard/blame/main/constants.py) — Defines word error rate and text normalization before scoring.
- [OpenMDW 1.1 license](https://openmdw.ai/license/1-1/) — License terms referenced by NVIDIA's model card.