Software maker Snowflake (SNOW.US) has launched a large-scale language model for enterprises called Snowflake Arctic.
The Zhitong Finance App learned that software manufacturer Snowflake (SNOW.US) has launched a large-scale language model for companies called Snowflake Arctic.
According to Snowflake, the company's open artificial intelligence model, called an “enterprise-grade LLM,” is optimized for complex enterprise workloads, surpassing multiple industry benchmarks in SQL code generation, instruction tracking, and more.
Snowflake stated that it will provide an LLM model with an Apache 2.0 license, which allows unrestricted personal, research, and commercial use.
Snowflake CEO Sridhar Ramaswamy said, “This is a watershed moment for Snowflake. Our AI research team is innovating at the cutting edge of artificial intelligence.”
For immediate use, Arctic can be used for serverless inference in Snowflake Cortex, a managed service that provides machine learning and AI solutions in a data cloud.
The company notes that the model will also be available on Amazon (AMZN.US) AWS, Hugging Face, Lamini, Microsoft (MSFT.US) Azure, Nvidia (NVDA.US) API catalog, Perplexity, Together AI, etc.
Snowflake says it also provides code templates and flexible inference and training options so users can quickly start customizing Arctic using their favorite frameworks. These include Nvidia NIM as well as Nvidia TensorRT-llm, vLLM, and Hugging Face.
In addition to Arctic LLM, the Snowflake Arctic series models include the recently released Arctic Embed, a series of text-embedding models.
The company notes that five models in the series are immediately available on the Hugging Face and will soon be available as part of Snowflake Cortex.