Version history

Version 1.0.0-beta49, released 2025-09-03

New features

• Add Validation mode in FunctionCallingConfig

Documentation improvements

• Update the comment for ANY mode in FunctionCallingConfig.

Version 1.0.0-beta48, released 2025-08-28

New features

• Add API response for Gen AI Evaluation in Tuning
• Add disable_container_logging in v1beta1 api

Version 1.0.0-beta47, released 2025-08-14

New features

• Add option EndpointUserId and ModelUserId fields
• Migrate dedicated endpoint to be enabled by default

Documentation improvements

• A comment for field dedicated_endpoint_enabled in message .google.cloud.aiplatform.v1beta1.DeployRequest is changed

Version 1.0.0-beta46, released 2025-08-13

New features

• Add API for Gen AI Evaluation in Tuning
• Add gpu_partition_size to MachineSpec

Version 1.0.0-beta45, released 2025-08-12

New features

• Added document_name for vertex ai search as part of retrieved context from grounding chunk
• Expose RecommendSpecs api to vertex python SDK for Custom Weights Model deployment

Version 1.0.0-beta44, released 2025-08-08

New features

• Add Grounding with Google Maps tool
• Add expiration for TTL for Memory and Sessions
• Add direct_memories_source
• Add force_delete field to DeleteRagFile 'preview' API request for Vertex RAG

Documentation improvements

• Update comment for allowed values for config models
• A comment for field monitored_resource_labels in message .google.cloud.aiplatform.v1beta1.AutoscalingMetricSpec is changed

Version 1.0.0-beta43, released 2025-08-05

New features

• Add embedding_metadata to google.cloud.aiplatform.v1.Index
• A new value NVIDIA_GB200 is added to enum AcceleratorType
• Add exclude_domains for grounding with GoogleSearch and EnterpriseWebSearch
• Online Prediction DeployModel API to support custom metrics based autoscaling
• Add FeatureViewDirectWrite API in v1

Version 1.0.0-beta42, released 2025-08-04

New features

• Add field ReasoningEngineSpec.service_account
• Add field ReasoningEngine.encryption_spec
• Add DeploymentStage for CreateEndpointOperationMetadata and DeployModelOperationMetadata
• Add the VeoTuningSpec
• Add enable_datapoint_upsert_logging to google.cloud.aiplatform.v1.DeployedIndex
• Add a FooBar API

Documentation improvements

• Update MutateDeployedModel documentation

Version 1.0.0-beta41, released 2025-07-28

New features

• Adds DWS and spot VM feature support to custom batch predictions 2.0
• Add service_account to Reasoning Engine public protos

Version 1.0.0-beta40, released 2025-07-22

Bug fixes

• Remove gemini_template_config and request_column_name fields from DatasetService.AssessData and DatasetService.AssembleData

Version 1.0.0-beta39, released 2025-07-15

New features

• Vertex AI Model Garden custom model deploy Public Preview
• Add API for Managed OSS Fine Tuning

Documentation improvements

• A comment for field model in message .google.cloud.aiplatform.v1beta1.TunedModel is changed
• A comment for field learning_rate_multiplier in message .google.cloud.aiplatform.v1beta1.SupervisedHyperParameters is changed
• A comment for field training_dataset_uri in message .google.cloud.aiplatform.v1beta1.SupervisedTuningSpec is changed
• A comment for field validation_dataset_uri in message .google.cloud.aiplatform.v1beta1.SupervisedTuningSpec is changed

Version 1.0.0-beta38, released 2025-07-11

New features

• Add flexstart option to v1beta1
• Some comments changes in machine_resources.proto to v1beta1
• Add Aggregation Output in EvaluateDataset Get Operation Response

Documentation improvements

• A comment for message DedicatedResources is changed
• A comment for field machine_spec in message .google.cloud.aiplatform.v1beta1.DedicatedResources is changed
• A comment for field min_replica_count in message .google.cloud.aiplatform.v1beta1.DedicatedResources is changed
• A comment for field max_replica_count in message .google.cloud.aiplatform.v1beta1.DedicatedResources is changed
• A comment for field required_replica_count in message .google.cloud.aiplatform.v1beta1.DedicatedResources is changed
• A comment for field min_replica_count in message .google.cloud.aiplatform.v1beta1.AutomaticResources is changed
• A comment for field max_replica_count in message .google.cloud.aiplatform.v1beta1.AutomaticResources is changed
• A comment for field boot_disk_type in message .google.cloud.aiplatform.v1beta1.DiskSpec is changed
• Add constraints for AggregationMetric enum and default value for flip_enabled field in AutoraterConfig

Version 1.0.0-beta37, released 2025-07-07

New features

• Add message ColabImage, add field colab_image to NotebookSoftwareConfig
• Allow user input for schedule_resource_name in NotebookExecutionJob
• Expose task_unique_name in pipeline task details for pipeline rerun
• Add computer use support to tools

Documentation improvements

• A comment for message NotebookRuntime is changed
• A comment for enum value BEING_STARTED in enum NotebookRuntime.RuntimeState is changed
• A comment for message NotebookSoftwareConfig is changed

Version 1.0.0-beta36, released 2025-06-25

New features

• Add PSCAutomationConfig to PrivateServiceConnectConfig in service_networking.proto
• Add GenAiAdvancedFeaturesConfig to endpoint.proto
• Reasoning Engine v1beta1 subresource updates
• Add invoke_route_prefix to ModelContainerSpec in aiplatform v1beta1 models.proto

Documentation improvements

• Update dedicateEndpointDns documentation
• Clarify that the names for sessions and session_events are no longer required.

Version 1.0.0-beta35, released 2025-06-17

New features

• Add DnsPeeringConfig in service_networking.proto
• Add dns_peering_configs to PscInterfaceConfig
• Add Scaled tier for RagEngineConfig to v1beta, equivalent to Enterprise
• Add Unprovisioned tier to RagEngineConfig in v1beta1 that can disable RagEngine service and delete all data within the service

Documentation improvements

• Enterprise tier in RagEngineConfig, use Scaled tier instead.

Version 1.0.0-beta34, released 2025-06-13

New features

• Introduce RagFileMetadataConfig for importing metadata to Rag

Version 1.0.0-beta33, released 2025-06-09

New features

• Expose UrlContextMetadata API to v1beta1
• Add json schema support to structured output and function declaration
• A new field include_thoughts is added to message .google.cloud.aiplatform.v1.GenerationConfig.ThinkingConfig
• A new field thought_signature is added to message .google.cloud.aiplatform.v1.Part
• A new field thought is added to message .google.cloud.aiplatform.v1.Part
• Expose URL Context API to v1beta1

Bug fixes

• Upgrade gRPC service registration func

Documentation improvements

• Allow field thought to be set as input

Version 1.0.0-beta32, released 2025-05-28

New features

• Introduce RAG as context/memory store for Gemini Live API
• Add ImportIndex to IndexService

Documentation improvements

• A comment for message RagFileParsingConfig is updated.
• A comment for field global_max_embedding_requests_per_min in message .google.cloud.aiplatform.v1beta1.ImportRagFilesConfig is updated.

Version 1.0.0-beta31, released 2025-05-20

New features

• Add VertexAISearch.max_results, filter, data_store_specs options

Version 1.0.0-beta30, released 2025-05-19

New features

• Add encryption_spec to Model Monitoring public preview API
• Adding thoughts_token_count to v1beta1 client library

Version 1.0.0-beta29, released 2025-05-15

New features

• Add checkpoint ID to endpoint proto
• Add ANN feature for RagManagedDb
• Tuning Checkpoints API
• New field additional_properties is added to message .google.cloud.aiplatform.v1beta1.Schema
• Add EncryptionSpec for RagCorpus CMEK feature
• A new value NVIDIA_B200 & NVIDIA_H200_141GB is added to enum AcceleratorType

Documentation improvements

• Fix links and typos
• Remove comments for a non public feature

Version 1.0.0-beta28, released 2025-05-06

New features

• A new field system_labels is added to message google.cloud.aiplatform.v1beta1.DeployRequest

Documentation improvements

• update an outdated URL

Version 1.0.0-beta27, released 2025-04-30

New features

• add gemini_request_read_config field to AssembleDataRequest and AssessDataRequest in aiplatform v1beta
• allow customers to set encryption_spec for context caching
• new fields ref and defs are added to message .google.cloud.aiplatform.v1beta1.Schema

Documentation improvements

• fix comment for field_mapping field in GeminiTemplateConfig
• Deprecate election category HARM_CATEGORY_CIVIC_INTEGRITY

Version 1.0.0-beta26, released 2025-04-29

Bug fixes

• BREAKING CHANGERemoved support for session resource paths that do not include reasoning engine ( commit 3f9a249 )

New features

• Adding ThinkingConfig to v1beta1 client library ( commit fa3ad70 )
• Add Model Garden EULA(End User License Agreement) related APIs ( commit 3f9a249 )
• Model Registry Model Checkpoint API ( commit 3f9a249 )
• Deprecated EventActions.transfer_to_agent and replaced with EventActions.transfer_agent ( commit 3f9a249 )

Version 1.0.0-beta25, released 2025-04-14

New features

• Add UpdateRagEngineConfig rpc ( commit 5f27e9d )
• Add RagEngineConfig to specify RAG project-level config ( commit 5f27e9d )
• Add rag_managed_db_config to RagEngineConfig for specifying Basic or Enterprise RagManagedDb tiers ( commit 5f27e9d )
• Add global quota config to vertex rag engine api ( commit 0156095 )
• Add FeatureViewDirectWrite API ( commit 41b15c9 )
• Add Gen AI logging public preview API ( commit e0cca8f )

Version 1.0.0-beta24, released 2025-03-31

New features

• Add page spans in retrieved contexts from Vertex RAG Engine in aiplatform v1beta1 ( commit e03b1fc )
• Add example, example_store, and example_store_service protos ( commit e5da036 )
• Add session.proto and session_service.proto ( commit ddc4e6a )
• Add batch prediction assessments to multimodal dataset RPCs ( commit 6d65548 )
• Enable force deletion in ReasoningEngine v1beta1 ( commit 4acec61 )
• Add model_config field for model selection preference ( commit 6135f3e )

Documentation improvements

• A comment for field rag_files_count in message .google.cloud.aiplatform.v1beta1.RagCorpus is changed ( commit e03b1fc )
• A comment for field model_name in message .google.cloud.aiplatform.v1beta1.RagFileParsingConfig is changed ( commit e03b1fc )

Version 1.0.0-beta23, released 2025-03-24

New features

• Update multimodal evaluation (content_map_instance), rubric generation (rubric_based_instance, etc) and raw_output(raw_output, custom_output, etc) proto change in online eval API ( commit 8f0bb30 )
• Add a ExportPublisherModel API ( commit 364b4aa )
• Add AssessData and AssembleData RPCs to DatasetService ( commit 82b9306 )
• Add import result gcs sink to the import files API ( commit f7a477d )
• Add import result bq sink to the import files API ( commit f7a477d )
• Add env variables and agent framework to ReasoningEngineSpec in v1beta1 ( commit ed0a196 )

Documentation improvements

• A comment for field autorater_config in message .google.cloud.aiplatform.v1beta1.EvaluateDatasetRequest is changed ( commit 8f0bb30 )
• A comment for field gcs_source in message .google.cloud.aiplatform.v1beta1.EvaluationDataset is changed ( commit 8f0bb30 )
• Update comment for package_spec from required to optional in ReasoningEngineSpec . ( commit ed0a196 )
• Add deployment_spec and agent_framework field to ReasoningEngineSpec . ( commit ed0a196 )

Version 1.0.0-beta22, released 2025-03-17

New features

• Add VertexAISearch.engine option ( commit 93bab4b )
• Add function_call.id and function_response.id ( commit ee278c3 )

Version 1.0.0-beta21, released 2025-03-10

New features

• Add multihost_gpu_node_count to Vertex SDK for multihost GPU support ( commit a99144f )
• Allowing users to specify the version id of the Model Garden model ( commit 33f9df1 )
• Allowing users to choose whether to use the hf model cache ( commit 33f9df1 )

Version 1.0.0-beta20, released 2025-03-04

Bug fixes

• BREAKING CHANGERemove VertexAISearch.engine option ( commit 18d4598 )

Version 1.0.0-beta19, released 2025-03-03

Bug fixes

• BREAKING CHANGEAn existing google.api.http annotation http_uri is changed for method DeployPublisherModel in service ModelGardenService ( commit 3da797d )

New features

• A new field include_equivalent_model_garden_model_deployment_configs is added to message .google.cloud.aiplatform.v1beta1.GetPublisherModelRequest ( commit 3da797d )
• Add VertexAISearch.engine option ( commit 05449f3 )
• Add EnterpriseWebSearch tool option ( commit 50d3edf )

Documentation improvements

• A comment for field model in message .google.cloud.aiplatform.v1beta1.DeployPublisherModelRequest is changed ( commit 3da797d )

Version 1.0.0-beta18, released 2025-02-25

New features

• Add Model Garden deploy API ( commit 9c643e2 )

Version 1.0.0-beta17, released 2025-02-10

New features

• A new field response_id is added to message .google.cloud.aiplatform.v1.GenerateContentResponse ( commit 96edae6 )
• A new field create_time is added to message .google.cloud.aiplatform.v1.GenerateContentResponse ( commit 96edae6 )
• Add Notebooks Runtime Software Configuration ( commit 40f33a4 )
• Add RolloutOptions to DeployedModel in v1beta1 endpoint.proto, add additional Probe options in v1beta1 model.proto ( commit aeee7bc )

Documentation improvements

• A comment for field filter in message .google.cloud.aiplatform.v1beta1.ListNotebookRuntimeTemplatesRequest is changed ( commit 40f33a4 )
• A comment for field filter in message .google.cloud.aiplatform.v1beta1.ListNotebookRuntimesRequest is changed ( commit 40f33a4 )

Version 1.0.0-beta16, released 2025-02-06

New features

• EvaluateDataset API v1beta1 initial release ( commit b343f9f )

Version 1.0.0-beta15, released 2025-02-03

New features

• Add rag_files_count to RagCorpus to count number of associated files ( commit 39f1a59 )
• Model Registry Checkpoint API ( commit fc5de1c )

Version 1.0.0-beta14, released 2025-01-27

New features

• Remove autorater config related visibility v1beta1 ( commit 0f0bb7a )
• Enable UpdateFeatureMonitor in v1beta1 API version ( commit c1ca304 )
• Add per-modality token count break downs for GenAI APIs ( commit fcf3080 )
• Enable FeatureGroup IAM Methods in v1beta1 API version ( commit f74c733 )
• Add machine_spec, data_persistent_disk_spec, network_spec, euc_config, shielded_vm_config to .google.cloud.aiplatform.v1beta1.NotebookRuntime ( commit c1bc4e9 )
• Add retrieval_config to ToolConfig v1beta1 ( commit d12475e )

Documentation improvements

• Deprecate is_default in message .google.cloud.aiplatform.v1beta1.NotebookRuntimeTemplate ( commit c1bc4e9 )
• Deprecate service_account in message .google.cloud.aiplatform.v1beta1.NotebookRuntimeTemplate ( commit c1bc4e9 )
• Deprecate service_account in message .google.cloud.aiplatform.v1beta1.NotebookRuntime ( commit c1bc4e9 )

Version 1.0.0-beta13, released 2025-01-06

New features

• Add Model Garden deploy API ( commit bc4928d )
• Add a new thought field in content proto ( commit f267357 )

Documentation improvements

• A comment for field labels in message .google.cloud.aiplatform.v1beta1.PublisherModel is changed ( commit bc4928d )

Version 1.0.0-beta12, released 2024-12-12

New features

• Add LLM parser proto to API ( commit 239b1c1 )
• Add Tool.GoogleSearch ( commit ec1d79d )
• Add GenerationConfig.Modality ( commit ec1d79d )
• Add GenerationConfig.SpeechConfig ( commit ec1d79d )
• Add GenerationConfig.MediaResolution ( commit ec1d79d )
• Enable FeatureGroup Service Account and IAM methods ( commit 6ee5a6b )
• A new value NVIDIA_H100_MEGA_80GB is added to enum AcceleratorType ( commit 0ced559 )
• A new field list_all_versions to ListPublisherModelsRequest ( commit afd7734 )

Version 1.0.0-beta11, released 2024-12-06

New features

• Add workbench_runtime and kernel_name to NotebookExecutionJob ( commit 83da91f )
• Add new Status field to DeployedModel in Endpoint ( commit 5846fb6 )
• Add new RequiredReplicaCount field to DedicatedResources in MachineResources ( commit 5846fb6 )
• Support streaming and multi class methods in Reasoning Engine v1beta1 API ( commit 6bbb600 )
• Introduce HybridSearch and Ranking configuration for RAG ( commit 15ac66e )
• Introduce VertexAiSearch integration for RAG ( commit 15ac66e )
• Trajectory eval metrics added to evaluation service proto ( commit 1c70f59 )
• Add CustomEnvironmentSpec to NotebookExecutionJob ( commit 92e884c )

Documentation improvements

• A comment for field encryption_spec in message .google.cloud.aiplatform.v1beta1.NotebookExecutionJob is changed ( commit 83da91f )
• A comment for field class_method in message .google.cloud.aiplatform.v1beta1.StreamQueryReasoningEngineRequest is changed (from steam_query to stream_query) ( commit d2f716f )
• Added support for multiple class_methods in QueryReasoningEngine ( commit 6bbb600 )
• Added support for StreamQueryReasoningEngine ( commit 6bbb600 )
• A comment for field api_key_config in message .google.cloud.aiplatform.v1beta1.JiraSource is changed ( commit 15ac66e )
• A comment for field partial_failure_gcs_sink in message .google.cloud.aiplatform.v1beta1.ImportRagFilesConfig is changed ( commit 15ac66e )
• A comment for field partial_failure_bigquery_sink in message .google.cloud.aiplatform.v1beta1.ImportRagFilesConfig is changed ( commit 15ac66e )
• A comment for field rag_file_parsing_config in message .google.cloud.aiplatform.v1beta1.ImportRagFilesConfig is changed ( commit 15ac66e )
• A comment for field source_uri in message .google.cloud.aiplatform.v1beta1.RagContexts is changed ( commit 15ac66e )
• Fixed typo for field use_strict_string_match in message .google.cloud.aiplatform.v1beta1.ToolParameterKVMatchSpec ( commit 1c70f59 )

Version 1.0.0-beta10, released 2024-11-18

New features

• Add a nfs_mounts to RaySpec in PersistentResource API ( commit 33bbebd )
• Add BYOSA field to tuning_job ( commit ce4d5a1 )
• Add BatchCreateFeatures rpc to feature_registry_service.proto ( commit d17404c )
• Add OptimizedConfig for feature_view ( commit d17404c )
• Add Feature Monitoring API to Feature Store ( commit e986f24 )
• Add StopNotebookRuntime method ( commit efa9ded )

Documentation improvements

• Update feature creation message commentary ( commit e986f24 )

Version 1.0.0-beta09, released 2024-10-30

New features

• Add UpdateEndpointLongRunning API in v1beta1 version ( commit 42c90e3 )

Version 1.0.0-beta08, released 2024-10-29

New features

• Add fast_tryout_enabled to FasterDeploymentConfig message in aiplatform v1beta1 endpoint.proto ( commit 546a95b )
• Add new PscInterfaceConfig field to custom_job.proto ( commit ce7c3b6 )

Version 1.0.0-beta07, released 2024-10-21

New features

• Add audio_timestamp to GenerationConfig ( commit dd3c974 )
• Introduce DefaultRuntime to PipelineJob ( commit 273b397 )

Documentation improvements

• A comment for field partner_model_tuning_spec in message .google.cloud.aiplatform.v1beta1.TuningJob is changed ( commit 2d1b3a7 )

Version 1.0.0-beta06, released 2024-10-07

New features

• Add enable_secure_private_service_connect in service attachment ( commit d90dd21 )
• Add new PscInterfaceConfig field to pipeline_job.proto ( commit 92d7924 )

Version 1.0.0-beta05, released 2024-09-30

Bug fixes

• Annotate PipelineJob and PipelineTaskRerunConfig fields as optional ( commit 7b00bf9 )

New features

• Add a dynamic retrieval API ( commit 76aa5a6 )

Version 1.0.0-beta04, released 2024-09-26

New features

• Add TunedModelRef and RebaseTunedModel Api for Vertex GenAiTuningService ( commit 7810c1b )
• Add CIVIC_INTEGRITY category to SafetySettings for prediction service ( commit 1071ade )
• Add psc_automation_configs to DeployIndex v1beta1 ( commit 9dad2ff )
• Add new PipelineTaskRerunConfig field to pipeline_job.proto ( commit b748a67 )
• Expose RuntimeArtifact proto in ui_pipeline_spec.proto ( commit b748a67 )
• A new field generation_config is added to message .google.cloud.aiplatform.v1beta1.CountTokensRequest ( commit 5723ee2 )
• A new field labels is added to message .google.cloud.aiplatform.v1beta1.GenerateContentRequest ( commit 5723ee2 )

Documentation improvements

• Limit comment SupervisedTuningSpec for 1p tuning ( commit 7810c1b )

Version 1.0.0-beta03, released 2024-09-16

New features

• A new field property_ordering is added to message .google.cloud.aiplatform.v1beta1.Schema ( commit 7ff096f )
• Add file_id to SharePointSource ( commit 536a4cc )
• Add share_point_sources to RagFile ( commit 536a4cc )
• Add Pinecone and Vector Search integration for Vertex RAG ( commit a63a2f8 )
• Add share_point_sources to SharePointSources ( commit eba1f55 )
• Add sharepoint_folder_path to SharePointSource ( commit eba1f55 )
• Add sharepoint_folder_id to SharePointSource ( commit eba1f55 )
• Add drive_name to SharePointSource ( commit eba1f55 )
• Add drive_id to SharePointSource ( commit eba1f55 )
• Add client_id to SharePointSource ( commit eba1f55 )
• Add client_secret to SharePointSource ( commit eba1f55 )
• Add tenant_id to SharePointSource ( commit eba1f55 )
• Add sharepoint_site_name to SharePointSource ( commit eba1f55 )
• Add share_point_sources to ImportRagFilesConfig ( commit eba1f55 )
• Add partial_failure_gcs_sink tp ImportRagFilesConfig ( commit eba1f55 )
• Add partial_failure_bigquery_sink to ImportRagFilesConfig ( commit eba1f55 )

Version 1.0.0-beta02, released 2024-09-09

New features

• Add OFF to HarmBlockThreshold ( commit c4c7995 )
• Add more configurability to feature_group.proto ( commit 82b085c )
• Add ragSource to feature_view.proto ( commit 82b085c )
• Add sync watermark to feature_view_sync.proto ( commit 82b085c )
• Add FLEX_START to Scheduling.strategy ( commit 6a5910f )
• Returns usage metadata for context caching ( commit beabf21 )
• Add Vector DB config for Vertex RAG (Weaviate + FeatureStore) ( commit 47a98d9 )
• Add UpdateRagCorpus API for Vertex RAG ( commit 47a98d9 )
• Add ApiKeyConfig field to ApiAuth ( commit 47a98d9 )
• Add max_wait_duration to Scheduling ( commit 71cd0dd )

Documentation improvements

• Fix typo in feature_online_store_admin_service.proto ( commit 82b085c )
• A comment for field vertex_prediction_endpoint in message .google.cloud.aiplatform.v1beta1.RagEmbeddingModelConfig is changed ( commit 47a98d9 )
• A comment for field distance in message .google.cloud.aiplatform.v1beta1.RagContexts is changed ( commit 47a98d9 )

Version 1.0.0-beta01, released 2024-08-14

Initial release.