Snowflake Arctic Embed

Snowflake Arctic Embed is an open-source family of text embedding models for developers building semantic search and document retrieval. It represents queries and documents as vectors so a search system can match them by meaning.

• Model size choices: Compact variants suit workloads where inference latency and cost matter, while larger variants prioritize retrieval accuracy.
• Longer passages: snowflake-arctic-embed-m-long accepts more text per passage than the standard models, making it useful for document collections that need longer context.
• Embedding API alternative: snowflake-arctic-embed-l provides an alternative to closed embedding APIs for document retrieval.

The smallest model, snowflake-arctic-embed-xs, builds on all-MiniLM-L6-v2. The small, medium and large variants use intfloat/e5 models, while the long-context variant builds on nomic-ai/nomic-embed-text-v1-unsupervised. The family uses the Apache 2.0 license.