Google Cloud → AWS and Azure
AWS and Azure equivalents of Gemini models on Agent Platform
The closest AWS equivalents of Google Cloud Gemini models on Agent Platform are Amazon Bedrock, Amazon Nova, Amazon Bedrock Data Automation, Amazon Rekognition and AWS HealthScribe; on Azure, Microsoft Foundry Models (incl. Azure OpenAI), Microsoft Foundry, Microsoft models in Foundry (MAI, Phi, Model Router), Azure Content Understanding, Azure Vision – Image Analysis, Azure AI Video Indexer, Azure Speech and Text Analytics for health.
About Gemini models on Agent Platform
| Provider | Google Cloud |
|---|---|
| Status | Generally available |
| What it is | Google's first-party models (Gemini, Gemma, Veo, Lyria, embeddings) served per token or per output unit through the Agent Platform API. |
| Documentation | docs.cloud.google.com |
| Pricing | Official pricing page |
Mappings (6)
Foundation model API platform Direct match
The Google Cloud equivalent of Amazon Bedrock is Gemini Enterprise Agent Platform (formerly Vertex AI): Gemini models plus partner models such as Claude and Grok offered as managed APIs through Model Garden.
- Vertex AI was renamed Gemini Enterprise Agent Platform on Apr 22, 2026. SDKs, IAM roles, and API hostnames still use the aiplatform / Vertex AI names.
- First-party models are Google's Gemini, Gemma, Veo, and Lyria. Bedrock's first-party family is Amazon Nova.
- Partner models (for example Anthropic Claude, xAI Grok, Mistral) are managed APIs in Model Garden. Open models (for example Llama, DeepSeek, Qwen) can be deployed to your own endpoints.
Cloud provider's own foundation models Direct match
Amazon Nova's Google Cloud counterpart is Google's own model family on Agent Platform: Gemini for text, multimodal, and reasoning, plus Veo for video, Lyria for music, and Gemma open models.
- Gemini models are multimodal by default (text, image, audio, video, and documents in one request).
- Gemma is an open-weights family you can also self-host, which Nova has no direct counterpart for.
- Both vendors retire model versions on published schedules. Check the model lifecycle page before pinning a version.
GenAI multimodal content extraction (IDP) Partial match
Bedrock Data Automation has no single counterpart. Document AI covers document extraction, and Gemini's multimodal input handles images, audio, and video with prompts you write.
- Document AI is document-focused (OCR, layout, forms, custom extractors). Audio and video extraction are done by prompting Gemini.
- Bedrock Data Automation gives one blueprint-driven API across documents, images, audio, and video.
Prebuilt image analysis Direct match
Amazon Rekognition Image maps to the Cloud Vision API (labels, objects, text, faces, landmarks, logos, SafeSearch), with Gemini for open-ended image understanding.
- Cloud Vision detects faces and facial attributes but does not identify individuals.
- Gemini can answer free-form questions about images, which neither pretrained API does.
Video analysis & indexing Partial match
Rekognition Video maps to the Video Intelligence API, but that API is deprecated (shutdown Sep 14, 2027). Google's forward path is Gemini, which accepts video input.
- Video Intelligence API was deprecated on Sep 14, 2026 and shuts down Sep 14, 2027.
- Gemini returns prompted descriptions and timestamps rather than fixed label/shot/object schemas.
Ambient clinical documentation (AI scribe) Partial match
AWS HealthScribe has no direct counterpart. Combine Speech-to-Text medical models (medical_conversation) with Gemini to draft clinical notes.
- You build the summarization prompts, evidence links, and safety checks yourself.
- Speech-to-Text medical models are in the v1 API.