Skip to main content
On Ona Cloud, Ona provides OpenAI model access for Codex Agent by default, with no configuration required. Customer-managed runners require OpenAI model access through Ona Intelligence or an approved OpenAI-compatible provider.
Ona Agent and Ona-managed Anthropic model access are no longer available on Ona Cloud. Use Codex Agent with OpenAI model access for new sessions and automations.
Use Ona-managed model access unless your Enterprise organization has an approved requirement to bring its own provider credentials. To see how much your organization uses, query the Ona Intelligence usage API for Ona-managed models, or the AI cost usage API if you bring your own provider keys (BYOK).

​
Codex Agent model access

Codex Agent is available on Ona Cloud and supported customer-managed runners. On Ona Cloud, Codex Agent can run using the model access available to the organization. Core users can also connect their ChatGPT account so Codex model requests use their ChatGPT plan. On a customer-managed runner, Codex Agent uses OpenAI model access provided by Ona Intelligence or an OpenAI-compatible LLM integration configured for that runner. See the provider guides for OpenAI direct API, AWS Bedrock (OpenAI API), and AWS Bedrock Mantle. See Codex Agent for how Codex works in Ona. To use your ChatGPT plan for Codex model requests, see Connect Codex to ChatGPT.

​
GPT-6.1 Sol

Use GPT-6.1 Sol with Ona-managed OpenAI access, a direct OpenAI API integration, AWS Bedrock Runtime, or AWS Bedrock Mantle. Your provider credentials must have access to the model. On Bedrock, access also depends on the AWS account and region. The dashboard offers Low, Medium, High, and Extra high reasoning, subject to organization policy. Direct OpenAI access supports Standard and optional Fast speed at 2x Standard cost. Bedrock supports Standard speed only; Fast speed and Ultra reasoning are unavailable. GPT-6.1 Sol requires a compatible Codex Agent installed in the environment. Starting a new environment does not upgrade the agent in an existing environment. Existing model defaults and saved selections remain unchanged. The Azure OpenAI setup does not offer deployment mappings for GPT-6.1 Sol. See Choose a model for selection and Troubleshooting if the model is missing.

​
GPT-6 Sol and Luna

Use GPT-6 Sol and GPT-6 Luna with Ona-managed OpenAI access, a direct OpenAI API integration, AWS Bedrock Runtime, or AWS Bedrock Mantle. Your provider credentials must have access to the selected model. On Bedrock, both models support Standard speed and Low through Max reasoning. Fast speed and Ultra reasoning are unavailable. Bedrock support requires an updated Codex Agent with Sol and Luna routing installed in the environment. The Azure OpenAI setup does not offer deployment mappings for Sol or Luna. See Choose a model for selection and reasoning options. Organization policies and the environment’s installed agent version also determine availability.

​
Custom token providers

Custom token providers are available by exception for Enterprise customers. If you have a requirement to use your own token provider, contact sales to discuss. For approved custom provider setups, see the provider-specific guides linked above. The IAM-based AWS Bedrock integration for Anthropic models, Google Vertex AI, and direct Anthropic API integrations are deprecated and maintenance-only.