Deploy DeepSeek R1-70B on GCP on Northflank

DeepSeek is a groundbreaking LLM which achieves a significant breakthrough in inference speed over previous models.

It tops the leaderboard among open-source models and rivals the most advanced closed-source models globally.

This stack template creates a new cluster in your own Google Cloud Platform account, and deploys Ollama and Open WebUI to run and query the DeepSeek R1:70B model in your own cloud account.

This template will deploy:

• 1 GCP cluster in the us-central1 region with:
  • 1 node pool of 3x n2-standard-4 nodes
  • 1 node pool of 1x a3-highgpu-1g nodes with 1 x NVIDIA H100 GPU
• A DeepSeek project with:
  • 1 Ollama deployment with a persistent volume
  • 1 Open WebUI deployment with a persistent volume

1. Create an account on Northflank
2. Integrate your GCP account
3. Click deploy Deploy DeepSeek on GCP now
4. Select your GCP account integration
5. Click deploy stack to save and run the DeepSeek template
6. When the template run finishes, view the open-webui service
7. Open the code.run domain and create a user account
8. Select the DeepSeek model and begin using your AI deployment!