Neo4j Aura Database December Release Update!

The Neo4j Aura December release is now rolling out—starting with AuraDB Free and gradually extending to higher-tier instances.

🤖 GenAI Plug-in

We have expanded our GenAI capabilities with native functions and procedures for major providers (e.g. OpenAI, Azure OpenAI, Google VertexAI, and Amazon Bedrock):

Vector Embeddings
: Use the new
ai.text.embed
function and
ai.text.embedBatch
procedure to generate Neo4j vector types directly from your data.
Text Completion
: Generate text from prompts using the new
ai.text.completion
function.
Provider Management
: Easily list allowed providers and configuration requirements with
ai.text.embed.providers
and
ai.text.completion.providers
.

🚀 New Cypher Features and Performance Improvements

Collection Functions
: We have expanded the Cypher 25 capabilities with a suite of new collection-based functions, including
coll.distinct
,
coll.flatten
,
coll.sort
,
coll.insert
, and more to give you finer control over your data structures.
Faster Dynamic Labels
:
MATCH
and
MERGE
queries utilizing dynamic labels (e.g.,
MATCH (p:$("Person"))
) now leverage property indexes, resulting in significantly faster performance.
Optimized Merge Operations
: We’ve introduced two new planner operators:
MergeUniqueNode
to speed up merge queries with uniqueness constraints, and
MergeInto
to improve performance on pattern creation queries.

🐞 Security and Bug Fixes

Security
: Resolved CVE-2025-12738 to tighten permission on property setting on and CVE-2025-11602 on Bolt protocol.
Fixed bugs related creation of full-text indexes, corrected error code 42N68 to 42I58 and fixed
ArrayIndexOutOfBoundsException
when using Label, Type, or property names longer than 2000 characters.

For full details of all updates and fixes in this release, please visit: Release Notes: Neo4j Aura Database – December 2025.