Snowflake Launches Top AI Model, Joins Open-Source AI on Hugging Face

Snowflake has launched a new AI model named 'snowflake-arctic-embed', which is touted as the world's best practical text-embedding model for retrieval use cases, surpassing competitors like OpenAI and Cohere. This launch is part of a broader move by Snowflake to enhance its AI capabilities, including state-of-the-art AI features integrated within its secure and governed environment, in collaboration with RekaAILabs. Additionally, Snowflake has joined the open-source AI community on Hugging Face, indicating a strategic shift towards more accessible AI technologies. The Snowflake AI team is continuing to develop small-sized, state-of-the-art (SOTA) embedding models that are also Apache-2.0 licensed. Support for these models is now available out-of-the-box in llama_index through Hugging Face integration, enabling users to build a RAG indexing pipeline over their data.