Skip to main content
The cooperbench config command provides interactive configuration wizards for setting up execution backends like GCP, Modal, and Docker.

​
Usage

cooperbench config <backend> [options]

​
Backends

​
GCP (Google Cloud Platform)

Configure Google Cloud Platform as your execution backend. 
cooperbench config gcp
This launches an interactive wizard that:
  1. Checks prerequisites - Verifies gcloud CLI and Python dependencies are installed
  2. Authenticates - Guides you through gcloud auth login and Application Default Credentials setup
  3. Configures project - Prompts for your GCP project ID and sets it as default
  4. Sets region/zone - Configures compute region and zone for task execution
  5. Validates setup - Tests API access to ensure everything works

​
Options

​
--skip-tests
flag
Skip validation tests after configuration. Faster setup, but skips verification that your GCP setup works correctly.

​
Example: Quick setup

cooperbench config gcp --skip-tests
This runs the configuration wizard but skips the API validation step at the end.

​
Example: Full setup with validation

cooperbench config gcp
The wizard will:
  • Check that gcloud CLI is installed
  • Verify Python GCP dependencies (google-cloud-batch, google-cloud-compute, google-cloud-storage)
  • Authenticate with both gcloud auth login and gcloud auth application-default login
  • Prompt for your project ID (or use current gcloud project)
  • Set compute region (default: us-central1) and zone (default: us-central1-a)
  • Test Compute Engine API access
  • Save configuration to ~/.config/cooperbench/config.json

​
Configuration storage

Configuration is saved to: 
~/.config/cooperbench/config.json
Example contents: 
{
  "gcp_project_id": "my-project-123",
  "gcp_region": "us-central1",
  "gcp_zone": "us-central1-a",
  "gcp_bucket": "cooperbench-eval-my-project-123"
}

​
Prerequisites

​
For GCP backend

1. Install gcloud CLI: 
brew install google-cloud-sdk
2. Install GCP Python dependencies: 
uv add --optional gcp 'cooperbench[gcp]'
Or with pip: 
pip install 'cooperbench[gcp]'
3. Enable required APIs: After configuration, enable these APIs in your GCP project: 4. Set up billing: Ensure your GCP project has billing enabled.

​
After configuration

Once configured, use the GCP backend with: 
cooperbench run --backend gcp
cooperbench eval --backend gcp
See Backends for more details on backend selection.