openclaw - 💡(How to fix) Fix Feature request: supported route for ChatGPT Pro / Deep Research style agents [1 participants]

It would be really useful to have a supported OpenClaw route for ChatGPT Pro style deep research / Research Pro workflows.

Right now I can use openai-codex/gpt-5.5 successfully through the Codex/OpenAI subscription path, but there does not seem to be a valid model route for a higher tier research mode. I tried configuring a dedicated research-pro agent with openai-codex/gpt-5.5-pro, but that model does not appear in openclaw models list, and the subagent runs completed with no usable output.

What I am looking for:

• A documented way to expose a true ChatGPT Pro / Deep Research / Research Pro style model or workflow inside OpenClaw, if the upstream provider supports it
• Clear model catalog naming so users do not guess routes like openai-codex/gpt-5.5-pro
• A recommended agent config pattern for high-value research agents that should be read-only, cite sources, and avoid file/message mutations
• If this is not currently possible through the Codex subscription/OAuth path, a note in the docs would help prevent confusion

Why this matters:

OpenClaw already makes it easy to create specialized agents. A dedicated research agent is one of the highest-value use cases, but it is hard to know whether the current limitation is model catalog support, provider access, subscription restrictions, or runtime behavior.

A good outcome would be either:

1. A supported model ID / provider route for this kind of Pro research workflow, or
2. Documentation saying it is not exposed yet, with the recommended fallback setup

Happy to test any suggested route or config.