Llama-Embed-Nemotron-8B Model Tops the Multilingual Text Retrieval Leaderboard

NVIDIA Nemotron continues to set the standard in retrieval and reranking performance, achieving outstanding results on the MMTEB multilingual text retrieval leaderboard.

The Llama-Embed-Nemotron-8B model—embodying Nemotron’s commitment to openness through shared datasets, weights, and training recipes—now ranks as the top open and portable model on MMTEB and is perfect for a wide range of text embedding tasks like retrieval, re-ranking, semantic similarity, classification, and bi-text mining.

Get started:

📺 Watch our video to learn more about how Nemotron leads the way in multilingual text retrieval.

🖥️ Check out Llama-Embed-Nemotron-8B on Hugging Face.

📖 Read our blog on Hugging Face to learn more about the model, architectural highlights, training methodology, performance evaluation and more.

📤 Stay up to date on NVIDIA Nemotron by subscribing to NVIDIA news and following NVIDIA AI on LinkedIn, X, YouTube and the Nemotron channel on Discord.