List of supported embeddings models
| Provider |
Model |
Multimodal |
Description |
Environment Support |
| openai |
text-embedding-ada-002
text-embedding-3-small
text-embedding-3-large |
No
No
No
No |
Must set dimensions to 1536, other values are not enabled by default(1) |
Beta, Production |
| azureopenai |
* |
|
Same models as openai, configure the associated URL endpoints |
Beta, Production |
| google/googlevertexai/vertex_ai(3) |
multimodalembedding@001(2)
text-embedding-004
text-embedding-005
text-multilingual-embedding-002
gemini-embedding-001(5) |
Yes
No
No
No
No |
The dimension parameter must be configured to 768(1) or 1408(2) |
Beta, Production |
| awsbedrock |
amazon.titan-embed-image-v1
cohere.embed-english-v3
amazon.titan-embed-v2
cohere.embed-english
cohere.embed-multilingual
amazon.nova-2-multimodal-embeddings-v1(6) |
Yes
Yes
No
Yes
No
Yes |
Check the model dimensions configuration(1,6) |
Beta, Production |
| nvidia |
baai.bge-m3
nv-embed-v1
nvclip
embed-qa-4
nv-embedqa-e5-v5
snowflake.arctic-embed-l
nv-embedqa-mistral-7b-v2 |
No
No
Yes
No
No
No
No |
Check the model dimensions configuration(1) |
Beta, Production |
| cohere |
embed-english-v3.0
embed-multilingual-v3.0
embed-english-light-v3.0
embed-multilingual-light-v3.0
embed-v4.0(4) |
Yes
No
No
No
Yes |
1024 dimensions except for the "light" models using 384, embed-v4.0 supports 256, 512, 1024, 1536 (default) |
Beta, Production |
| pinecone(2) |
multilingual-e5-large
llama-text-embed-v2 |
No
No |
1024 dimensions. |
Beta |
| mistral |
mistral-embed
codestral-embed |
No
No |
1024 dimensions.
1536 dimensions by default (maximum value: 3072) |
Beta |
(1) - For On-premises installations, the associated module must be installed and configured.
(2) - The configuration must be done in the Retrieval tab by configuring the Profile Metadata parameter. For more information, refer to Not registered Providers and Models. Please note that the associated Embeddings costs are not recorded, and no related Requests are logged as they are processed by the provider.
(3) - Check the Provider recommendations.
(4) - Supported since 2025-07 Release.
(5) - Supported since 2025-08 Release; the dimension parameter defaults to 1536, other options are available
(6) - Supported since 2026-01 Release; output dimension options are 3072, 1024, 384, and 256.