sub-agent work SUPER SLOW

What version of the Codex App are you using (From “About Codex” dialog)?

ChatGPT+Codex App Version 26.707.31123

What subscription do you have?

Codex Pro

What platform is your computer?

Microsoft Windows NT 10.0.26200.0 x64

What issue are you seeing?

Sub agent work using GPT 5.6 Sol xHigh is incredibly slow. Takes forever to get it up and running, then it takes forever, burns through tokens, hangs up and the controlling agent has to poke it multiple times to check to see if ti's still running. Pretty much unusable at the moment. I had tasks that should take no more than 5-10 minutes and it takes 30.

I have not checked to see if the problem is the same in other models or intelligences. Token burn is too high and i have work to do.

What steps can reproduce the bug?

Running a single or multi-sub-agent run on ChatGPT GPT 5.6 Sol xHigh.

What is the expected behavior?

It should run without all the hangups and slow starts.

Additional information

please fix soon and please allow us a way to check the output or debug of the sub-agents to see the token spend.