OpenAI Model Connector
The OpenAI Model Connector defines OpenAI as the LLM provider and default model for your Worker Agents.
What you will learn
- Which OpenAI models the connector supports.
- What each reasoning effort level means for cost and latency.
- How to configure authentication, the default model, and reasoning effort.
- How the Harness-managed connector offering works if you do not have model access.
Supported models
The OpenAI Model Connector supports GPT-5.5 with configurable reasoning effort.
Reasoning effort levels
The reasoning effort level controls how much reasoning the model applies before responding, trading latency and cost against depth. Set it on the connector configuration form.
| Effort level | Description |
|---|---|
low | Fastest, lowest-cost responses for simple tasks |
medium | Balanced reasoning and latency for most tasks |
high | Deeper reasoning for complex tasks |
xhigh | Maximum reasoning depth for the most demanding tasks |
Connector configuration
When you create the connector, set the following:
- Authentication: Provide an OpenAI API key. Reference the key from a Harness secret rather than entering it in plain text.
- Model Name: Select the default model (GPT-5.5) the agent uses at runtime.
- Reasoning Effort: Select the default effort level (
low,medium,high, orxhigh).
Harness-managed connector
If you do not have access to a model provider, Harness offers a managed LLM connector you can use instead.
Until August 2026, usage of the Harness-managed LLM connector is included in your Harness subscription at no additional cost. After August 2026, Harness bills managed LLM connector usage separately, in addition to your Harness subscription.
Related concepts
- Go to Worker Agents to create and configure agents that use this connector.
- Go to Anthropic Model Connector to configure Claude as your model provider.