Releasesby GlyphRay-2.56.0
Highlights Ray Data Stability: In this Ray release, we've added a variety of stability improvements, including running multiple datasets in a cluster, adding automatic batch size selection to CPU-based map-batches, and default logical memory configuration to prevent OOMs.
Also shipped Jun 29, 2026Anyscale in June 2026 →
Dynamic Iceberg table replication (General availability)
Snowflake now supports replication of dynamic Iceberg tables as part of replication and failover groups. Previously, dynamic Iceberg tables were skipped during replication refresh operations. For more information, see Configure replication for Snowflake-managed Iceberg tables .
Databricks SQL version 2026.20 is now available in Preview
Databricks SQL version 2026.20 is now available in the Preview channel. Review the following section to learn about new features, behavioral changes, and bug fixes.
SQL alert task in Lakeflow Jobs is now generally available
The SQL alert task in Lakeflow Jobs is now generally available. Use the task to evaluate a Databricks SQL alert as part of a Lakeflow Job. The task returns its evaluation state as a task output value, so downstream tasks can branch on the result. See SQL alert task for jobs .
Govern LLMs in Unity Catalog with model services (Beta)
You can now govern Databricks-hosted LLMs in Unity Catalog using model services , currently in Beta. A model service is a Unity Catalog securable that represents a governed LLM endpoint, so you can define an endpoint once and share it across workspaces using Unity Catalog privileges, instead of duplicating endpoints per workspace.
Custom Docker images for AI Runtime CLI workloads (Beta)
You can now bring your own Docker container image to AI Runtime CLI ( air ) workloads. This feature is in Beta . Use a custom image when you need specific system library versions, complex dependencies, or a reproducible environment. See Use custom Docker images .
Parquet v2 for Delta Lake tables is generally available
Parquet v2 is now generally available for Delta Lake tables in Databricks Runtime 18.1 and above. Parquet v2 uses advanced encodings, v2 data page headers, and INT64 timestamps to improve query performance and reduce table sizes. To use, set the delta.parquet.format.version table property to 2.12.0 or enable automatic upgrades.
More from RayEvery Ray release →
Ray-2.57.0
Highlights Ray Data: In this Ray release, we've enabled DataSourceV2 by default ( #64821 ), so read_parquet and friends use the new scan/listing infrastructure with row-group-aware chunking and predicate splitting. Hash Shuffle V2 eliminates the aggregator actor pool.
Ray-2.56.1
Highlights Ray Data : We added fixes for several to_pandas regressions introduced in 2.56: an opt-out flag ( RAY_DATA_ENABLE_ARROW_BACKED_PANDAS_CONVERSION ) for Arrow-backed conversion, an int64/ double[pyarrow] overflow crash on concatenation, and a TensorDtype.__from_arrow__ crash on empty tensor columns ( #64793 , #64794 ).
Ray-2.58.0
Highlights Ray Serve LLM: In this release we've completed KV cache and token aware request routing, which was previewed in 2.57. Tokenization now happens in-process on the LLMRouter ingress replica, the routing decision is made there, tokens are transmitted out-of-band so the engine does not re-tokenize, KV lifecycle events are broadcast to every ingress...
Ray-2.55.1
- Fixes SSH connectivity issue in the ray-llm image ( #62625 / #62718 ). Upgrade apt packages in slim base ( #62666 / #62717 ).
Ray-2.55.0
Ray Data 🎉 New Features Add DataSourceV2 API with scanner/reader framework, file listing, and file partitioning ( #61220 , #61615 , #61997 ) Support GPU shuffle with rapidsmpf 26.2 ( #61371 , #62062 ) Add Kafka datasink, migrate to confluent-kafka , support datetime offsets ( #60307 , #61284 , #60909 ) Add Turbopuffer datasink ( #58910 ) Add 2-phase commit...
Ray-2.59.0
Highlights > * Ray Data LLM & Ray Serve LLM are GA/Stable: the LLM APIs graduate to general availability this release (\#65194), alongside an upgrade to vLLM 0.27.0 (\#65351).
Ray-2.54.1
Ray Data 🔨 Fixes Disable hanging issue detection ( #61895 ) , The hanging issue detector was making blocking calls to the Ray State API, which could cause the scheduling loop to block and severely degrade pipeline performance. The detector is disabled in this patch release until the blocking calls are fixed.
Ray-2.58.1
Fixes Ray Data - Fix out-of-scope block callbacks being registered on blocks passed through from input to output (\#66369, cherry-pick \#66404). Fixes spurious callback errors on passthrough datasets. Images - google-cloud-cli is now installed in the base-extra images (\#65757, pick \#66611). Contributors Thanks to everyone who contributed to this release!