Account-specific issue with the ChatGPT Desktop app for macOS
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 125k
- Forks
- 19.4k
- PR merge metrics
- PR metrics pending
Description
What version of the Codex App are you using (From “About Codex” dialog)?
ChatGPT 26.901.51231
What subscription do you have?
Owner of a ChatGPT Business workspace.
What platform is your computer?
No response
What issue are you seeing?
Hello,
I’m experiencing a persistent account-specific issue with the ChatGPT Desktop app for macOS.
Problem
I am the Owner of a ChatGPT Business workspace.
On the ChatGPT web app, my account works correctly and I can select different reasoning levels, including Medium, High, and Extra High.
In the ChatGPT Desktop app, however:
- Chat mode is permanently stuck on “Instant”.
- I can move the reasoning/intelligence slider, but the displayed selection remains Instant.
- Higher reasoning levels do not actually get selected or persist.
- The model picker itself appears to load, and models such as GPT-5.6 Sol and GPT-5.5 are visible.
- Work mode works completely normally, including its model and reasoning controls.
Important A/B test
This appears to be specific to my account/session in Desktop Chat.
An employee who is only a Member of the same Business workspace logged into the same ChatGPT Desktop installation on the same MacBook, and higher reasoning levels in Chat mode worked normally for him.
Therefore: - Same Mac → works for another account
- Same Desktop app → works for another account
- My account on Web Chat → works
- My account in Desktop Work → works
- Only my account in Desktop Chat is affected
Environment - macOS
- ChatGPT Desktop: 26.901.51231
- Build: 8109
- Apple Silicon / arm64
- Bundle ID: com.openai.codex
- ChatGPT Business workspace
- Role: Owner
I have already verified that the application installation, local state, VS Code/Codex environment, and related local databases are structurally healthy. A controlled reset of only the persisted Desktop Chat model-picker selection had no effect, so this does not appear to be caused by a stale local model selection.
The issue appears more consistent with an account-specific Desktop Chat model policy/entitlement response, rollout assignment, or Desktop renderer reconciliation issue.
Could you please check whether my account/workspace has an incorrect Desktop Chat model/reasoning entitlement, feature flag, rollout assignment, slider_settings, or workspace_model_policy?
The key point is that the same account has the correct reasoning options on the web, but not in Desktop Chat.
What steps can reproduce the bug?
Feedback ID: no-active-thread-01a0828f-e4ae-7222-82fb-c73d7c92595a
What is the expected behavior?
No response
Additional information
No response
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reproducing the account-specific behavior in Desktop Chat and comparing it with the same account on the web and in Desktop Work, as well as the member account on the same Mac. Investigate the entitlement, feature-flag, rollout, or workspace-policy response and the Desktop renderer reconciliation mentioned in the report. Done means the account can select and retain higher reasoning levels in Desktop Chat.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- macos
- Domain
- desktop
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100