Timeout when starting a new conversation (newConversation); switching model works #4541
Description
What version of Codex is running?
codex 0.4.15
Which model were you using?
all
What platform is your computer?
Windows vscodex
What steps can reproduce the bug?
A new conversation starts within a few seconds without errors.
The CLI reports a timeout and fails to create the local task when starting a new conversation.
Logs
2025-10-01 11:47:26.454 [warning] Request f81542dc-ecf1-441b-90d0-80bdc9223fba newConversation {"cwd":"","model":"gpt-5","config":{},"approvalPolicy":"never","baseInstructions":null,"sandbox":"danger-full-access","profile":null,"includePlanTool":true,"includeApplyPatchTool":null} timed out
2025-10-01 11:47:26.455 [error] Error creating local task: Error: Timeout
What is the expected behavior?
Workarounds
Switch to another model → chat resumes immediately (context preserved).
Delete the failed conversation and retry → works.
Repro Frequency
Intermittent; only happens at newConversation start. Once a session is active, continuing or switching models is fine.
Additional Notes
All models can be listed and used outside of newConversation.
Happy to provide verbose logs if there’s a flag (e.g., --verbose, DEBUG=*).
If large workspaces during initialization are known to cause delays, I can try ignoring node_modules, .git, .astro, etc., and report back.
What do you see instead?
No response
Additional information
No response