[TUI 0.145.0] Esc-Esc Edit/Resume changes gpt-5.6-sol Ultra to xhigh
Nobody has claimed this yet.
Assessment
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Newbie friendliness
- 68/100
Research direction
Start with the Rust TUI's Esc-Esc Edit/Resume flow and the prompt-edit branch creation or attachment path described in the reproduction. Trace how the active conversation's reasoning effort is carried into the branch, then verify that selecting Ultra remains Ultra after Esc-Esc and Enter.
Written by the indexing model from the issue text.
Description
What version of Codex CLI is running?
codex-cli 0.145.0
What subscription do you have?
ChatGPT Pro 5x
Which model were you using?
gpt-5.6-sol
What platform is your computer?
Microsoft Windows NT 10.0.26200.0 x64
What terminal emulator and version are you using (if applicable)?
Windows Command Prompt
Codex doctor report
Not collected. The issue reproduces deterministically in a short, otherwise normal TUI session.
What issue are you seeing?
After selecting gpt-5.6-sol with Ultra reasoning, using the Esc-Esc
Edit/Resume flow silently changes the reasoning effort back to xhigh.
Ultra remains selected during normal conversation turns. The change occurs
immediately after pressing Enter to confirm a previous prompt, before the
restored prompt is submitted again.
What steps can reproduce the bug?
-
Start Codex CLI 0.145.0.
The TUI initially displays
gpt-5.6-sol xhigh. -
Run
/model, selectgpt-5.6-sol, and selectUltra.The TUI now displays
gpt-5.6-sol ultra. -
Send
1+1=?.Codex answers
2, and the TUI still displaysgpt-5.6-sol ultra. -
Send
2+2=?.Codex answers
4, and the TUI still displaysgpt-5.6-sol ultra. -
Press Esc twice.
-
The previous prompt,
2+2=?, is highlighted. Press Enter to select it
for Edit/Resume. -
Before submitting the restored prompt, observe that the model indicator
has changed togpt-5.6-sol xhigh.
What is the expected behavior?
The prompt-edit branch should preserve the active conversation's selected
model and reasoning effort, so it should remain gpt-5.6-sol ultra.
The reasoning effort should not silently fall back to the configured
non-Ultra default when creating a branch from an earlier prompt. If this
reset is intentional, the TUI should clearly disclose it before creating
the branch.
Additional information
- The issue reproduces with one Esc-Esc-Enter gesture; repeated or rapid
input is not required. - No reasoning-effort shortcut keys were used.
- Normal conversation turns preserve
Ultra. - The visible change occurs during prompt-edit branch creation or attachment,
not after a new model response. - The observed change is confirmed by the TUI model indicator. The next
request payload was not independently inspected. - A possible cause is that the prompt-edit fork path uses the configured
default effort (xhigh) rather than copying the active conversation's
current effort (Ultra).
Related issues
- #35286 involves the same Esc-Esc-Enter prompt-edit fork path, but reports
buffered input creating multiple chained forks. This issue reproduces with
a single gesture and concerns the selected reasoning effort not being
preserved. - #35003 involves the same earlier-prompt fork feature in Codex CLI 0.145.0,
but reports copied record timestamps being rewritten. It does not report
a model or reasoning-effort change. - #11013 and #12042 report reasoning effort changing when entering Plan mode.
This reproduction does not enter Plan mode; it is triggered by confirming
an Esc-Esc Edit/Resume selection. - #26930 reports reasoning effort changing in Codex Desktop during
delegation or continuation. This report is a deterministic Codex CLI TUI
reproduction involving a prompt-edit branch.
- Dominant language
- Rust
- Stars
- 125k
- Forks
- 19.5k
- Avg merge
- 1m
- Merged PRs (30d)
- 1k
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.
More from openai/codex
-
enhancement remote
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
-
bug CLI windows-os
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
-
macOS sandbox blocks hw.optional.arm64 sysctl, causing Flutter to misdetect Apple Silicon as x64 Openbug CLI sandbox
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
-
bug CLI TUI
Difficulty 2/5 1-3 hours Newbie friendliness 90/100
-
CLI config enhancement skills
Difficulty 2/5 1-3 hours Newbie friendliness 84/100
Similar issues
-
Difficulty 2/5 1-3 hours Newbie friendliness 86/100
kwakseongjae/auto-hwp#319 ·
-
area:cli bug filter-quality good first issue priority:medium
Difficulty 2/5 1-3 hours Newbie friendliness 84/100
-
Difficulty 1/5 Under an hour Newbie friendliness 72/100
bevyengine/bevy#25861 ·
-
comp-datalake
Difficulty 2/5 1-3 hours Newbie friendliness 88/100
ClickHouse/ClickHouse#121222 ·
-
A-linter
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
oxc-project/oxc#26863 ·