Build with Jamba-Instruct on Snowflake Cortex

AI21 is integrating its Jamba-Instruct large language model with the Snowflake Data Cloud via the Snowflake Cortex managed service. This partnership provides Snowflake customers with secure and streamlined access to build generative AI applications. Jamba-Instruct features a groundbreaking hybrid architecture combining Transformer and SSM models, which enables a massive 256K context window. This large context capacity allows for powerful, cost-efficient enterprise use cases like sophisticated RAG engines and analysis of long documents.