Releasesby GlyphSpanner supports expression indexes on scalar expressions
Secondary indexes can be built on scalar expressions over one or more columns, removing the need for dedicated generated columns just for indexing.
Also shipped Aug 28, 2026Google Cloud in August 2026 →
Monitor and cap external provider costs
Unity Gateway can now track and cap spend on external model providers, such as Amazon Bedrock and Azure AI Foundry, not just Databricks-provided models. If your organization already has capacity commitments directly with a provider, Unity Gateway becomes the single point to track that usage and enforce spend limits across teams and use cases.
Unity Catalog Skills (Beta)
You can now publish, govern, and share agent skills with Unity Catalog. A skill is a first-class Unity Catalog securable ( catalog.schema.skill ) that follows the SKILL.md specification , so you register it once and control who can discover, use, and update it with the same permissions, tags, and audit that protect your tables.
Hy4 Preview now available on AI Gateway
Hy4 Preview from Tencent is now available on AI Gateway. Hy4 Preview is an open-source Mixture-of-Experts model with 770B total parameters and 49B active per token, aimed at long-horizon coding, document analysis, game development, and scientific reasoning. It serves a context window of 1M tokens.
GitHub Copilot weekly releases , August 24
This week’s updates give you more control over how Copilot runs, from team sessions in Slack and Teams to customization across the app, CLI, and your IDE. GitHub Copilot in Slack and Microsoft Teams Turn team conversations into shared agent sessions.
Upcoming changes to GitHub Copilot policies and billing
To provide a strong, consistent Copilot experience, we’re making three separate, upcoming changes to Copilot policies and billing. Please review the upcoming updates to understand what may impact you.
Training Serverless Training deprecation: Qwen 3.5 9B and Qwen 3.6 27B Qwen 3.5 9B and Qwen 3.6 27B are...
Training Serverless Training deprecation: Qwen 3.5 9B and Qwen 3.6 27B Qwen 3.5 9B and Qwen 3.6 27B are deprecated from Serverless Training effective August 26, 2026. Migrate new and existing Serverless Training workloads to Qwen 3.8 27B . This change applies to the shared Serverless Training pool.
More from SpannerEvery Spanner release →
Spanner PostgreSQL dialect supports the TABLESAMPLE operator
PostgreSQL-dialect Spanner databases can use TABLESAMPLE to select a random sample of a dataset.
Spanner PostgreSQL dialect adds ILIKE and NOT ILIKE
PostgreSQL-dialect databases in Spanner now support the ILIKE and NOT ILIKE operators and the pg.ilike and pg.not_ilike functions for case-insensitive pattern matching.
Spanner Graph adds semantic options and key uniqueness validation
CREATE PROPERTY GRAPH can set description and synonyms options on labels and properties to help natural language query interfaces, and a validate_element_key_uniqueness option controls validation at schema creation.
Spanner applies the 80,000 mutation limit per DML statement
The DML mutation limit now applies per statement instead of across the whole transaction, so transactions can run several DML statements that together exceed it. The 100 MiB commit size limit still applies.
Spanner dynamic channel pooling for Go and Java clients is GA
Dynamic channel pooling for gRPC channels prevents issues from under- or over-provisioned channels and reduces configuration overhead in the Go and Java client libraries. It is disabled by default.
Spanner queues are generally available (GA)
Spanner queues are generally available (GA) . Spanner queues provide transactional messaging to help you manage asynchronous work. The feature pairs this capability with the scalability and reliability of Spanner, letting you build event-driven applications. For more information, see the Spanner queues overview .
Spanner adds automatic parameterization of SQL query literals
Spanner converts hardcoded literals in CRUD-style queries such as primary key lookups, index lookups and key joins into parameters, so execution plans are cached and reused to cut latency and CPU costs.
Spanner supports creating tables without defining primary keys
Spanner supports creating tables without defining primary keys. When you create a table without a primary key, Spanner creates a hidden rowid column that serves as the primary key. For more information, see Create a table without defining a primary key .