2 people reported

All time, since Oct 16.

Early automatic context compaction

“Is automatic context compaction too aggressive under the new context window limit?”

Is this your bug? What gets counted gets fixed. Pile on.

Paste this into your agent. It files your report, with your error and version.

I hit this bug: rapsheet.dev/sheets/codex/issues/18288/file
Follow it to file my report. Show me the text before you send.

No install. No account. Public and anonymous. The link asks your agent for one report, nothing else. What it sends

Auto-compaction starts well below displayed 475k context window. Compaction begins while the context indicator shows "66% used" (312k / 475k tokens), or after "200-something-k tokens", leaving part of the window unused.

Is this your bug?

File my report
Where
Context compaction
Platforms
Windows
How bad
Breaks a normal workflow.

Not quite your bug?

  1. Premature auto-compaction26 people
  2. Context window shows wrong size4 people
  3. Compaction and re-inspection loop3 people

Original reports2, newest first

  1. [Windows desktop] GPT-6.1 Sol sometimes auto-compacts at ~200k usage despite a displayed 475k context windowOn GitHub#55008
  2. Automatic context compaction starts at only 66% usage (312k / 475k tokens)On GitHub#54517

Do you work on Codex?

Claim its RAP Sheet to see what your users report and answer them in one place.