9.40 Release Notes: Dec 15, 2025-Jan 09, 2026

Attention

This release has completed. For differences between the in-advance and final versions of these release notes, see Release notes change log.

New features

Notifications for data quality incidents (Preview)

Snowflake can automatically send a notification when there is a data quality incident in a database. A data quality incident occurs when the return value of a data metric function (DMF) violates an expectation or constitutes an anomaly.

Notifications can be sent via email or through an external system like Slack, Teams, and PagerDuty.

For more information, see Sending notifications for data quality issues.

Deprecated features

Deprecation of external OpenAI model routing for Cortex Analyst

Snowflake has deprecated the ENABLE_CORTEX_ANALYST_MODEL_AZURE_OPENAI account parameter that routes Cortex Analyst requests to external OpenAI GPT models using Azure OpenAI outside the Snowflake secure perimeter.

Snowflake deprecated this legacy configuration for the following reasons:

  • It sends data outside Snowflake’s trusted environment.

  • It uses older model versions.

  • Newer GPT and Claude models now run fully within the Snowflake secure perimeter, offering better text-to-SQL accuracy and stronger security.

Snowflake no longer honors this parameter.

Cross-region inference (optional)

We strongly recommend enabling cross-region inference to access the full set of LLMs:

ALTER ACCOUNT SET CORTEX_ENABLED_CROSS_REGION = 'ANY_REGION';
Copy

Enabling this parameter allows Snowflake to serve requests from the best available LLMs, which might reside in another Snowflake region, while keeping all processing within the Snowflake secure perimeter.

If you prefer to keep inference within your current region, the LLMs must be available in-region. If your region does not have LLMs available, you must enable cross-region inference to use Cortex Analyst.

For more information about the models used by Cortex Analyst, see Control models used by Cortex Analyst.

SQL updates

Semantic views: Using standard SQL clauses to query semantic views (Preview)

You can now use standard SQL clauses in a SELECT statement to query a semantic view.

This feature is in Preview.

You can just specify the name of the semantic view in the FROM clause, rather than specifying the SEMANTIC_VIEW clause. For example, the following query specifies the SEMANTIC_VIEW clause:

SELECT * FROM SEMANTIC_VIEW(
    tpch_analysis
    DIMENSIONS customer.customer_market_segment
    METRICS orders.order_average_value
  )
  ORDER BY customer_market_segment;
Copy

The following statement demonstrates how to execute the same query without specifying the SEMANTIC_VIEW clause:

SELECT customer_market_segment, AGG(order_average_value)
  FROM tpch_analysis
  GROUP BY customer_market_segment
  ORDER BY customer_market_segment;
Copy

For information, see Specifying the name of the semantic view in the FROM clause.

Release notes change log

Announcement

Update

Date

Release notes

Initial publication (preview)

Dec 15, 2025

Semantic views: Using standard SQL clauses to query semantic views (Preview)

Added to SQL updates

Dec 16, 2025

Copy tags when running a CREATE OR REPLACE TABLE command (Preview)

Removed from Data governance updates

Jan 06, 2026

Deprecation of external OpenAI model routing for Cortex Analyst

Added to Deprecated features

Jan 06, 2026