---
title: "Granite Embedding Multilingual R2: Open Apache 2.0 Multilingual Embeddings with 32K Context — Best Sub-100M Retrieval Quality"
description: "Granite Embedding Multilingual R2 introduces two Apache 2.0 multilingual embedding models built on ModernBERT: a 97M-parameter compact model and a 311M full-size model. Both cover more than 200 languages, receive enhanced retrieval training for 52 languages and nine programming languages, and support context lengths up to 32,768 tokens, compared with 512 for the R1 models. The compact model scores 60.3 on MTEB Multilingual Retrieval, while the full-size model scores 65.2; the reported results place them among the strongest open models in their parameter ranges. R2 replaces the XLM-RoBERTa architecture with alternating attention lengths, rotary position embeddings, newer multilingual and code tokenizers, and Flash Attention 2.0 support, while the 311M model adds Matryoshka embeddings and both models provide ONNX and OpenVINO weights."
---

# Granite Embedding Multilingual R2: Open Apache 2.0 Multilingual Embeddings with 32K Context — Best Sub-100M Retrieval Quality

[Hugging Face](https://yomu.fyi/company/hugging-face) · Radu Florian, Parul Awasthy, Aashka Trivedi, Madison Lee · May 14, 2026

**Type:** Announcement

## Summary

Granite Embedding Multilingual R2 introduces two Apache 2.0 multilingual embedding models built on ModernBERT: a 97M-parameter compact model and a 311M full-size model. Both cover more than 200 languages, receive enhanced retrieval training for 52 languages and nine programming languages, and support context lengths up to 32,768 tokens, compared with 512 for the R1 models. The compact model scores 60.3 on MTEB Multilingual Retrieval, while the full-size model scores 65.2; the reported results place them among the strongest open models in their parameter ranges. R2 replaces the XLM-RoBERTa architecture with alternating attention lengths, rotary position embeddings, newer multilingual and code tokenizers, and Flash Attention 2.0 support, while the 311M model adds Matryoshka embeddings and both models provide ONNX and OpenVINO weights.

## Context

Multilingual embedding systems face a trade-off between broad language coverage and model size: smaller models are faster but may sacrifice languages or retrieval quality. The release addresses multilingual retrieval, cross-lingual search, retrieval-augmented generation over multilingual corpora, and code retrieval across international teams.

## Approach / What changed

The R2 models use ModernBERT encoders, multilingual and code-capable tokenizers, retrieval-pair and cross-lingual training for 52 languages, and training data assembled from IBM-curated, public, internally generated, and synthetic sources. The 311M model uses knowledge distillation, contrastive fine-tuning, model merging, and Matryoshka Representation Learning; the 97M model uses a pruned 180K-token GPT-OSS-derived vocabulary and a pruning methodology. Both ship with ONNX and OpenVINO weights.

## Takeaways

- The 97M model scores 60.3 on MTEB Multilingual Retrieval across 18 languages, compared with 50.9 for multilingual-e5-small, according to the post.
- The 311M model produces 768-dimensional embeddings and supports Matryoshka truncation, allowing a storage and speed trade-off; the 97M model produces 384-dimensional embeddings.
- Both models support more than 200 languages, enhanced retrieval quality for 52 languages, 32,768-token contexts, and cross-lingual code retrieval across Python, Go, Java, JavaScript, PHP, Ruby, SQL, C, and C++.

**Tags:** [Machine Learning](https://yomu.fyi/topic/machine-learning), [Open Source](https://yomu.fyi/topic/open-source), [Performance](https://yomu.fyi/topic/performance), [Search](https://yomu.fyi/topic/search)

- Source: [Hugging Face](https://huggingface.co/blog/ibm-granite/granite-embedding-multilingual-r2)
- Source URL: https://huggingface.co/blog/ibm-granite/granite-embedding-multilingual-r2
- Ingested by Yomu: 2026-08-29T01:48:59.408Z

[Read original post](https://huggingface.co/blog/ibm-granite/granite-embedding-multilingual-r2)