Supported Models
Embedding Models
Section titled “Embedding Models”| Model | Dimensions | Languages | Output Types |
|---|---|---|---|
| BAAI/bge-m3 | 1024 | 100+ | dense, sparse, multivector |
| BAAI/bge-large-en-v1.5 | 1024 | en | dense |
| BAAI/bge-base-en-v1.5 | 768 | en | dense |
| BAAI/bge-small-en-v1.5 | 384 | en | dense |
| dunzhang/stella_en_1.5B_v5 | 8192 | en | dense |
| dunzhang/stella_en_400M_v5 | 8192 | en | dense |
| Alibaba-NLP/gte-Qwen2-1.5B-instruct | 1536 | multi | dense |
| intfloat/e5-large-v2 | 1024 | en | dense |
| intfloat/e5-base-v2 | 768 | en | dense |
| intfloat/multilingual-e5-large | 1024 | 100+ | dense |
| nomic-ai/nomic-embed-text-v1.5 | 768 | en | dense |
| sentence-transformers/all-MiniLM-L6-v2 | 384 | en | dense |
Reranking Models
Section titled “Reranking Models”| Model | Max Length | Languages |
|---|---|---|
| BAAI/bge-reranker-v2-m3 | 8192 | 100+ |
| BAAI/bge-reranker-large | 512 | en |
| BAAI/bge-reranker-base | 512 | en |
| cross-encoder/ms-marco-MiniLM-L-6-v2 | 512 | en |
Extraction Models
Section titled “Extraction Models”| Model | Task |
|---|---|
| urchade/gliner_multi-v2.1 | Named entity extraction |
Model Selection
Section titled “Model Selection”For general use: Start with BAAI/bge-m3 — good quality, multilingual, all output types.
For English-only: BAAI/bge-large-en-v1.5 or intfloat/e5-large-v2.
For maximum quality: dunzhang/stella_en_1.5B_v5 (larger, English-only).
For speed: BAAI/bge-small-en-v1.5 or sentence-transformers/all-MiniLM-L6-v2.