Aleph Alpha's open-weight English-German mixture-of-experts model for self-hosting, released under Apache 2.0 in October 2026
Nemotron 3 Diarization
New· 4NVIDIA's open-weight speaker diarization model, which labels who spoke when for up to eight speakers, offline or streaming
About Nemotron 3 Diarization
Transcription tells you what was said. Diarization works out who said it. For that second job NVIDIA offers Nemotron 3 Diarization, a model with open weights that arrived on 23 September 2026 with a model card, a write-up on Hugging Face and press coverage.
How the model works
Input is 16 kHz mono audio. At 10-millisecond intervals the model outputs, for up to eight speakers, the chance that each one is speaking, twice the four speakers of NVIDIA's earlier Streaming Sortformer. It works offline on full recordings or in streaming mode, with latency chosen from 0.32, 0.64 and 1.04 seconds, the kind of setting live meeting and call-center products need. With roughly 100 million parameters it is light enough to sit beside a speech-recognition model rather than replace one.
Languages and licensing
The model card lists English, Mandarin, Hindi, Kannada, Telugu and Bengali. NVIDIA's blog says training data spanned 21 languages while evaluation used English, so performance in other languages deserves testing before anyone depends on it. Commercial use is permitted under the OpenMDW 1.1 license. For hosted access, NVIDIA points to Baseten, DigitalOcean and Argmax, and it sets no usage price itself.
Reported accuracy
On Voice Arena's independent Diarization-Bench, which uses 139 far-field English sessions totaling about 22 hours and is labeled early results, the model ranked first of twelve systems with a 14.72 percent diarization error rate when read on 25 September 2026. Two other systems it ranked ahead of scored 19.34 and 20.56 percent. NVIDIA's separate figures include 12.73 percent on DIHARD III and a slight step back on two-speaker CALLHOME; those come from NVIDIA itself.
TrustList summary
Released on 23 September 2026, Nemotron 3 Diarization is NVIDIA's open-weight model for telling speakers apart in audio. Diarization is the step in a transcription pipeline that works out who spoke when. The model has about 100 million parameters, small enough to run alongside a speech-recognition model.
What Nemotron 3 Diarization offers
- Speaker probabilities: every 10 milliseconds, gives the probability that each of up to eight speakers is talking, doubling the four-speaker limit of NVIDIA's earlier Streaming Sortformer.
- Offline or streaming: works offline on full recordings, or live with latency settings of 1.04, 0.64 or 0.32 seconds.
- Audio input: takes 16 kHz mono audio.
- Hosted routes: NVIDIA names Baseten, DigitalOcean and Argmax as hosted options.
Clients
Streaming suits live meeting and call-centre products, and the offline mode suits recorded audio.
Key facts
- Licence: OpenMDW 1.1, which allows commercial use.
- Languages: English, Mandarin, Hindi, Kannada, Telugu and Bengali on the model card. NVIDIA says training covered 21 languages and evaluation was in English.
- Benchmark: first of twelve systems on Voice Arena's Diarization-Bench, at a 14.72 per cent error rate, labelled early results and read on 25 September 2026.
- Developer's figures: 12.73 per cent on DIHARD III, per NVIDIA.
Summary written by TrustList from the business's own published information, Nov 2026.
Benchmarks & AI stats
| Benchmark | Official | Community avg |
|---|---|---|
| No benchmark scores yet. Be the first to add one. | ||
“Official” values are editor-approved and feed the ranking. “Community avg” is the mean of member submissions (shown for transparency; it never affects the ranking until an editor approves a value).
Sign in to add a benchmark score for this model.
Nemotron 3 Diarization launches
Trust Score
An earned signal from verification, reviews, awards, transparency and engagement - the vendor can't buy it.
Updated 9/25/2026
Request a demo or quote from Nemotron 3 Diarization
Reviews
Write the first review of Nemotron 3 Diarization
Used it? Your experience helps other buyers decide.
Write a reviewQuestions & answers
No questions yet. Be the first to ask about Nemotron 3 Diarization.
Nemotron 3 Diarization alternatives
Similar listings buyers compare against this one.
Liquid AI's open-weight decision model, giving yes or no, choice or score answers from text and images
Black Forest Labs' image generation and editing model with bounding-box layouts, multiple reference images and 4K output
Microsoft AI's real-time speech-to-text model for voice agents, transcribing 60 languages with continuous language detection
Perplexity's open-weight decision model that picks among given options for routing and checks in AI systems
Microsoft AI's text-to-speech model for keeping one voice across 23 languages, with voice cloning from short audio
Anthropic's smallest and fastest Claude 5.5 model, available on its own API and three cloud platforms
Google's open multimodal embedding model that maps text, images, audio and video into one vector space, under Apache 2.0
More on TrustList
Everything here links back to the same verified catalogue. Pick your next stop.
- Write a reviewHelp other buyers decide
- CompaniesAgencies, consultancies and IT service providers, ranked by verified reviews.
- ProductsSoftware and SaaS with pricing, features, integrations and alternatives.
- ArticlesGuides, comparisons and research from the editorial desk and the community.
- AwardsAnnual recognition decided by verified reviews and an independent jury.
- LaunchesNew products and releases, voted up by the community every day.
- RequestsBuyers describe what they need; vendors respond directly.
- PeopleReviewers, authors and makers with public profiles.
- ComparePut up to four listings side by side before you shortlist.