AI agents · AWS vs Google Cloud
Amazon Bedrock AgentCore vs Agent Runtime (formerly Agent Engine)
Amazon Bedrock AgentCore maps to Agent Runtime (formerly Agent Engine) on Agent Platform, which hosts and scales agents with sessions, Memory Bank, code execution, and observability.
Side-by-side comparison
| Aspect | Amazon Web Services | Google Cloud |
|---|---|---|
| Service | Amazon Bedrock AgentCore | Agent Runtime (formerly Agent Engine) |
| Lifecycle | Generally available | Renamed from Vertex AI Agent Engine on Apr 22, 2026. |
| What it is | Modular agent platform. Harness (managed agent loop), Runtime, Memory, Gateway (MCP), Identity, Code Interpreter, Browser, Observability, Evaluations, Optimization, Policy, Registry, Payments. Works with any framework and any model. | Managed runtime to deploy and scale agents built with ADK or other frameworks, with Sessions, Memory Bank, code execution, and observability. |
| Pricing model | Consumption-based per AgentCore component (for example Runtime compute while active, Gateway calls, Memory events/records, Browser and Code Interpreter sessions), plus model tokens. No upfront commitment. | Runtime is billed for vCPU-hours and GiB-hours of memory used. Sessions, Memory Bank, and code execution are billed separately, plus model tokens. |
| Docs | Amazon Bedrock AgentCore docs | Agent Runtime (formerly Agent Engine) docs |
| Pricing page | Amazon Bedrock AgentCore pricing | Agent Runtime (formerly Agent Engine) pricing |
Key differences
- Agent Runtime was renamed from Vertex AI Agent Engine on Apr 22, 2026.
- Both are framework-agnostic. Google's first-party framework is ADK; AWS's is Strands Agents.
- Code Execution and Computer Use run in sandboxes inside Agent Runtime, comparable to AgentCore Code Interpreter and Browser.
Pricing summary
Pricing models only. No prices are quoted because they change by region and over time.
- AWS
- Consumption-based per AgentCore component (for example Runtime compute while active, Gateway calls, Memory events/records, Browser and Code Interpreter sessions), plus model tokens. No upfront commitment. Amazon Bedrock AgentCore pricing →
- GCP
- Runtime is billed for vCPU-hours and GiB-hours of memory used. Sessions, Memory Bank, and code execution are billed separately, plus model tokens. Agent Runtime (formerly Agent Engine) pricing →
And on Azure?
Direct match Azure's equivalent is Foundry Agent Service. AWS vs Azure comparison →
FAQ
What is the Google Cloud equivalent of Amazon Bedrock AgentCore?
Amazon Bedrock AgentCore maps to Agent Runtime (formerly Agent Engine) on Agent Platform, which hosts and scales agents with sessions, Memory Bank, code execution, and observability.
How does pricing compare between Amazon Bedrock AgentCore and Agent Runtime (formerly Agent Engine)?
AWS: Consumption-based per AgentCore component (for example Runtime compute while active, Gateway calls, Memory events/records, Browser and Code Interpreter sessions), plus model tokens. No upfront commitment. Google Cloud: Runtime is billed for vCPU-hours and GiB-hours of memory used. Sessions, Memory Bank, and code execution are billed separately, plus model tokens. Check the official pricing pages for current rates.
What is the Azure equivalent?
Foundry Agent Service (direct match). The Azure equivalent of Amazon Bedrock AgentCore is Foundry Agent Service.