Vertex AI Prediction Agent Platform Inference
Legacy names Vertex AI Prediction, Vertex AI Inference, Vertex AI Batch Prediction, Vertex AI Batch Inference, Vertex AI Online Prediction, Vertex AI Online Inference, Vertex AI Endpoints, AI Platform Prediction
Managed online (real-time endpoint) and asynchronous batch model serving and inference service.
Naming timeline
- AI Platform Prediction
Managed model serving service succeeding Cloud ML Engine online/batch prediction on ml.googleapis.com.
- Vertex AI Prediction
Unified online endpoint serving and batch prediction jobs on aiplatform.googleapis.com.
- Vertex AI Inference
Documentation and console labels shifted from 'Vertex AI Prediction' to 'Vertex AI Inference' (Online Inference and Batch Inference).
- Agent Platform InferenceCurrent name
Rebranded to Agent Platform Inference ('Gemini Enterprise Agent Platform Inference'), covering Agent Platform Online Inference, Agent Platform Batch Inference, and Agent Platform Endpoints.
What is Vertex AI Prediction called now?
Vertex AI Prediction is now called Agent Platform Inference, with the latest change in October 2025 (Gemini Enterprise Launch (Release Notes Table Apr 22, 2026)). Earlier names covered by this entry also include Vertex AI Inference, Vertex AI Batch Prediction, Vertex AI Batch Inference, Vertex AI Online Prediction, Vertex AI Online Inference, Vertex AI Endpoints, AI Platform Prediction. Managed online (real-time endpoint) and asynchronous batch model serving and inference service. Only the customer-facing name changed; the underlying technical identifiers remain backward compatible.
Also known as
- Gemini Enterprise Agent Platform Inference
- Agent Platform Online Inference
- Gemini Enterprise Agent Platform Online Inference
- Agent Platform Batch Inference
- Gemini Enterprise Agent Platform Batch Inference
- Agent Platform Endpoints
- Gemini Enterprise Agent Platform Endpoints
Technical identifiers
Technical identifiers unchanged. Only the customer-facing product name changed; the APIs, CLI commands, IAM roles and Terraform resources below keep their legacy identifiers for backward compatibility.
| REST / gRPC APIs | aiplatform.googleapis.com |
|---|---|
| gcloud / CLI Commands | gcloud ai endpoints gcloud ai endpoints predict gcloud ai endpoints raw-predict |
| IAM Predefined Roles | roles/aiplatform.* |
| Terraform Resources | google_vertex_ai_endpoint |
Programmatic methods and resources retain 'predict'/'Prediction' (endpoints.predict, BatchPredictionJob, gcloud ai endpoints predict) on aiplatform.googleapis.com.
Notes & disambiguation
- Underwent a two-step terminology and brand change: first shifting from 'Vertex AI Prediction' (Online Prediction / Batch Prediction) to 'Vertex AI Inference' (Online Inference / Batch Inference), and then rebranding under Agent Platform to 'Agent Platform Inference' ('Gemini Enterprise Agent Platform Inference') and 'Agent Platform Endpoints'.
- Despite the customer-facing shift from 'Prediction' to 'Inference', underlying REST/gRPC APIs (BatchPredictionJob, endpoints.predict, endpoints.rawPredict), CLI commands ('gcloud ai endpoints predict'), and documentation URL paths (/machine-learning/predictions) retain 'predict' and 'prediction' identifiers.
Sources & official documentation
- Official documentation ↗
- Product page ↗
- Overview of Getting Inferences on Agent Platform — Google Cloud Documentation
- Gemini Enterprise Agent Platform Name Changes (Inference, Batch Inference, Online Inference & Endpoints)
- Gemini Enterprise Agent Platform Release Notes (April 22, 2026 Naming Changes Table)
Catalog last updated . Spotted an error? Suggest an edit