Announcing Claude Sonnet 5 on Snowflake Cortex AI
Snowflake announces the immediate availability of Anthropic’s Claude Sonnet 5 in private preview on Snowflake Cortex AI, expanding enterprise AI capabilities within its secure environment.
The useful question is what changes for users, developers or buyers, and whether the announcement stays industry context or becomes something people can actually use.
Snowflake has launched Anthropic’s latest model, Claude Sonnet 5, as a same-day private preview on Snowflake Cortex AI. The model is accessible within Snowflake’s secure perimeter for use with Snowflake CoCo, Cortex Agents, Cortex AI Functions, Cortex Inference, and Snowflake CoWork. Sonnet 5 demonstrates strong performance in coding, debugging, deep analysis, and brainstorming, with improved agentic capabilities compared to its predecessor. It supports long-horizon execution tasks and integrates better with persisted memory, enhancing workflows in environments like Snowflake CoCo.
Preview customers can leverage Sonnet 5 to accelerate development with Snowflake CoCo, Snowflake’s AI coding agent. The model enhances reasoning, code generation, and complex agentic tasks, approaching Opus-level capability at reduced cost. It provides Snowflake-aware assistance across CoCo Desktop, Snowsight interface, CLI, and local development tools. For data engineers and analysts, Sonnet 5 enables natural language conversion into production-ready pipelines, analytics, and AI agents, reducing manual work while maintaining governance and enterprise-grade interoperability.
Snowflake CoWork, a personal work agent for knowledge workers, now supports Sonnet 5 for analyzing structured and unstructured data with high precision. Cortex AI Functions allow building scalable AI pipelines across multimodal enterprise data using SQL, offering cost-effective processing compared to manual orchestration. Sonnet 5 improves vision tasks, including chart and table data extraction, and achieves Opus-level document parsing accuracy on financial and medical documents, particularly for dense content.
Developers can build agentic enterprise applications within Snowflake’s security perimeter using Cortex Inference and Sonnet 5’s adjustable effort levels, balancing reasoning depth against latency. The model includes self-verification loops for autonomous operation with greater consistency and reduced supervision. Integration is streamlined via the Anthropic SDK, pointing to the Snowflake Cortex Inference endpoint. Enterprises can adopt Sonnet 5 across multiple entry points, accelerating the deployment of production-ready agentic AI at scale.