EMBEDDINGS FOR SEARCH
GARNIgarni-embed-1
ԳառնիГарниGarni
Dense vectors for search, clustering and retrieval across languages, so a question asked in one language finds the answer written in another.
Best for
- Semantic search
- Knowledge bases and RAG
- Duplicate and similarity detection
- Document classification
Named after
Garni, the 1st-century Greco-Roman temple above the Azat river gorge.
Specification sheet
| API identifier | garni-embed-1 |
|---|---|
| Type | Embeddings |
| Stage | Generally available |
| Context window | 8K |
| Vector size | 1024 |
| Typical latency | ≈ 40 ms per batch of 32 |
| Primary region | AM-EVN-2 |
| Deployment | Cloud API · Private cloud · On-premises |
| Price | $0.02 per 1M tokens |