DeepSeek R1 Available in Google Vertex AI Model Garden

Unlocking cutting-edge AI capabilities just became easier with the arrival of DeepSeek R1 in Google Cloud’s Vertex AI Model Garden. This addition underscores Google Cloud's commitment to fostering an open, adaptable AI environment, equipping organizations with more choices to drive innovation and address unique business challenges.

Managed Model-as-a-Service: Simplifying AI Adoption

Deploying advanced models like DeepSeek R1 typically requires significant infrastructure investments, including specialized GPUs and technical expertise. Vertex AI’s Model-as-a-Service (MaaS) solution removes these hurdles by providing secure, fully managed, serverless APIs. 

Teams can access powerful AI without handling hardware or complex setup, shifting focus toward developing impactful applications. The platform guarantees robust security, privacy, and compliance, all through a scalable, pay-as-you-go approach.

Quick Start Guide: DeepSeek R1 in Vertex AI

1. Enable the DeepSeek API Service

Getting started is straightforward. Visit the DeepSeek API Service page in the Vertex AI Model Garden, open the model card, and enable the service. After activation, usually within minutes, you’re ready to use DeepSeek R1.

2. Experiment with the UI

The built-in interface lets you interact with DeepSeek R1 instantly. Use the sidebar in the model card to test sample queries or tasks, allowing you to assess the model’s performance before full integration.

3. Integrate via API

For development teams, Vertex AI supports both REST API calls and the OpenAI Python API Client Library. This flexibility accommodates diverse tech stacks and project needs. DeepSeek’s endpoint is secured and isolated, no outbound internet access, ensuring your data stays protected.

• REST API: Authenticate with gcloud credentials and send prediction requests from your terminal or application.
  
  
• Python Client: Install the OpenAI library, set up credentials and endpoint, and make chat completion requests programmatically.

Empowering Teams with Credits and Resources

To encourage experimentation, Google Cloud offers $300 in free credits for new users, plus ongoing free tiers for over 20 products, including AI APIs. This minimizes risk and upfront cost, making it accessible for teams to prototype and refine AI-driven solutions.

Explore, Build, and Connect

• Check out the DeepSeek R1 documentation for comprehensive setup and usage guidance.
• Review pricing details to plan your scaling strategy.
• Browse the Model Garden to discover more open-source models.
• Start with a proof-of-concept to evaluate DeepSeek R1’s fit for your business needs.
• Engage with the Google Cloud AI Community to share insights and learn from fellow practitioners.

Takeaway: Focusing on Innovation over Infrastructure

With managed services like DeepSeek R1, Vertex AI Model Garden removes technical obstacles and accelerates AI-driven transformation. Whether you’re prototyping or deploying at scale, these advancements let you focus on creating value letting the platform handle the complexity of infrastructure and security.

Source: Google Cloud Blog: DeepSeek R1 is available for everyone in Vertex AI Model Garden