Anthropic Claude Opus 5 launches on Snowflake Cortex AI

Snowflake announced the availability of Anthropic's Claude Opus 5 in public preview on Snowflake Cortex AI, with Snowflake acting as a launch partner providing same-day access to the model. Claude Opus 5 operates within Snowflake's secure perimeter and integrates with five core Snowflake AI products: CoCo (Snowflake's AI coding agent), Cortex Agents, Cortex AI Functions, Cortex Inference, and Snowflake CoWork.

Clause Opus 5 is positioned as a "thoughtful, proactive model" that achieves state-of-the-art performance on coding and knowledge-work evaluations while delivering improved performance over its predecessor, Opus 4.8. The model comes close to the frontier intelligence of Claude Fable 5 and shows meaningful gains on scientific-research tasks. A key feature is its adjustable effort settings, which allow users to optimize either for intelligence or for faster results and token conservation. The model is also described as stronger at verifying its own work and iterating until it succeeds, produces stronger visual outputs, and is well suited to demanding, long-horizon agentic and analytical work.

Several use cases are highlighted across Snowflake's product suite. CoCo Desktop, now generally available on macOS and Windows, enables developers to transform natural language into production-ready pipelines, analytics, and AI agents. Unlike general-purpose AI coding tools, CoCo comes with the user's Snowflake catalog already loaded and RBAC roles active, eliminating cold starts and governance gaps. Snowflake CoWork functions as a personal work agent for knowledge workers, using Deep Research to decompose complex business questions across structured and unstructured enterprise data and return fully cited answers in minutes, plus proactive automations that surface insights and recommended actions. Cortex AI Functions make it easy to build scalable AI pipelines using SQL, with the AI_COMPLETE function allowing analysts to apply natural language instructions directly. Cortex Inference enables developers to build agentic enterprise applications within the Snowflake security perimeter, with Opus 5's adjustable effort levels helping balance reasoning depth and latency per task.

Snowflake notes that enterprises want to move agentic AI into production quickly, and Opus 5 enables agents to handle complex, long-running tasks with greater consistency and reduced supervision. The multiple entry points—from CoCo CLI to CoWork, Cortex AI Functions, and Cortex Inference—give customers flexibility in how they adopt the capability. Developers can integrate Opus 5 using the Anthropic SDK by pointing it to their Snowflake Cortex Inference endpoint. The announcement includes a forward-looking statement noting that while the company is excited about future product offerings, actual results may differ due to known and unknown risks and uncertainties.

Snowflake's announcement of Claude Opus 5 availability represents a direct integration of Anthropic's latest frontier model into the Snowflake ecosystem. By securing same-day launch availability as a partner with Anthropic, Snowflake is positioning itself as a platform where enterprises can deploy cutting-edge AI without sacrificing the security and governance controls they require. The model's improvements over Opus 4.8—particularly in coding, knowledge work, and scientific research—address the specific needs of data engineers, analysts, and developers who work within enterprise environments.

The integration spans multiple Snowflake AI products, each targeting different workflows. CoCo Desktop, now generally available, focuses on AI-assisted code generation and development, collapsing "weeks of glue work into minutes" for pipeline and analytics work. Snowflake CoWork brings agentic reasoning to knowledge workers via Deep Research and proactive automations. Cortex AI Functions and Cortex Inference provide lower-level APIs for developers building custom applications. This multi-layer approach allows different roles within an enterprise to adopt Opus 5 without requiring a single unified interface or rearchitecture.

The inclusion of adjustable effort settings—allowing teams to trade intelligence for speed and token efficiency—suggests Snowflake and Anthropic recognize that not all tasks require frontier-level reasoning. This flexibility is particularly important for enterprises running long-horizon agentic tasks at scale, where cost and latency can constrain production deployments.

No comments yet. Be the first to share your thoughts!