Models

Models represent the configuration for a large-language model, such as OpenAI GPT and Claude Opus.

Ark agents can be configured to use any model and will use the model named default if a specific model is not specified.

You can quickly configure a model by running ark dashboard and going to the ‘Models’ page.

Creating Models

Create a Model Resource. As an example. this is how you would configure the an OpenAI gpt-4o-mini model:


# First, create a secret with your OpenAI API key.
# Secrets are the recommended way to store sensitive data.
kubectl create secret generic default-model-secret \
  --from-literal=api-key='YOUR_OPENAI_API_KEY'
 
# Create the model resource, which uses the configured secret for its API
# key.
cat <<EOF | kubectl apply -f -
apiVersion: ark.mckinsey.com/v1alpha1
kind: Model
metadata:
  name: default
spec:
  type: openai
  model:
    value: gpt-4o-mini
  config:
    openai:
      baseUrl:
        value: "https://api.openai.com/v1"
      apiKey:
        valueFrom:
          secretKeyRef:
            name: default-model-secret
            key: api-key
EOF

You can then query the model directly:


ark query model/default "Explain how AI large language models work"

You can also create a model with the command ark models create. For example, to create the default model use:


ark models create default

You will then be asked to provide the model configuration, API key, etc.

Ready to create agents? Move to the Agents guide.