Granite Embedding English R2
Granite Embedding English R2 is a 149M English retrieval model with 768-dimensional output. Released on 15 August 2025, it succeeds the earlier 125M English embedding model.
Model design and output
The model encodes queries and documents independently for semantic retrieval. Its 8,192-token limit applies to each encoded input. It is a retrieval component, so an application needs separate search and answer-generation steps.Granite Embedding English R2 model card
Practical use
Practical guidance. Use it for English collections where a dedicated retrieval encoder meets the language requirement. Check retrieval on short questions, full paragraphs and domain abbreviations. Chunk documents around meaningful sections so unrelated content does not dilute the vector.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | Downloadable |
| Model identifiers | ibm-granite/granite-embedding-english-r2 |
| Context window | 8,192 tokens Published maximum input length. Granite Embedding English R2 model card |
| License | Apache 2.0Granite Embedding English R2 model card |
| Architecture | ModernBERT bi-encoderGranite Embedding English R2 model card |
| Vector size | 768 dimensionsGranite Embedding English R2 model card |
Related models
- Granite Embedding Reranker English R2
IBM · Granite Reranker
A 149M ModernBERT cross-encoder that scores English query-document pairs to refine search results.
Launch date Downloadable - Granite Embedding Small English R2
IBM · Granite Embedding
A 47M ModernBERT English retrieval encoder that produces 384-dimensional vectors.
Launch date Downloadable