Release
NVIDIA releases Nemotron 3 Embed multilingual text embedding models in 1B and 8B sizes
NVIDIA's Hugging Face model cards give July 16, 2026 as the release date of Nemotron-3-Embed-8B-BF16 and Nemotron-3-Embed-1B-BF16, text embedding models for retrieval and semantic similarity in retrieval-augmented generation systems. The cards say the models were evaluated across 34 languages and output 4096-dimensional (8B) and 2048-dimensional (1B) vectors that can be shortened by slicing and re-normalizing. Both are licensed under OpenMDW-1.1, and the 8B card says it is built with Mistral AI's Ministral-3-8B-Instruct-2512, released under Apache 2.0. USASI has not tested the models.12
Related catalog entries
Sources
Share or cite this page
- X (opens X)
- Bluesky (opens Bluesky)
- LinkedIn (opens LinkedIn)
- Reddit (opens Reddit)
- Hacker News (opens Hacker News)
- Email (opens your email app)
- RSS updates
Cite this page
APA style
USASI. (2026, October 13). NVIDIA releases Nemotron 3 Embed multilingual text embedding models in 1B and 8B sizes. United States of America Superintelligence. https://unitedstatesofamericasuperintelligence.com/news/nvidia-releases-nemotron-3-embed/
BibTeX
@misc{usasi_news_nvidia_releases_nemotron_3_embed,
author = {{USASI}},
title = {NVIDIA releases Nemotron 3 Embed multilingual text embedding models in 1B and 8B sizes},
year = {2026},
month = oct,
howpublished = {\url{https://unitedstatesofamericasuperintelligence.com/news/nvidia-releases-nemotron-3-embed/}},
note = {United States of America Superintelligence. Last updated 2026-10-13}
}The date is when this page was last updated. Add the date you read it if your style needs one.