{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreiahv5a2o3ixzyoo4hxykt64z6uyrgyewsn5fktby56q3xavehdw6m",
"uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mgmqpcvadni2"
},
"path": "/t/codex-app-on-windows-became-extremely-slow-today-even-in-fast-mode/1376087#post_1",
"publishedAt": "2026-03-09T10:42:32.000Z",
"site": "https://community.openai.com",
"textContent": "Hi team, I’d like to report a performance issue with the Codex app on Windows.\n\nStarting today, Codex has become extremely slow for me. A single response can take more than 20 minutes, even for a small project.\n\nMy setup:\n\n * Codex app on Windows\n\n * Model: GPT-5.4\n\n * /fast mode enabled\n\n * Repository size: around 20 files\n\n * Project is on a Windows path\n\n\n\n\nWhat makes this unusual is that the same setup was working much faster before. The slowdown started suddenly today.\n\nWhat I already checked:\n\n * Opened a brand new thread\n\n * Kept the task small\n\n * /fast mode is already enabled\n\n * The repo is very small, so this does not seem related to repo size\n\n * Previously, the same Windows-path workflow was fast, so this does not look like a simple “Windows path is slow” issue\n\n\n\n\nObserved behavior:\n\n * Codex stays in “thinking” for a very long time\n\n * Even simple prompts take 20+ minutes\n\n * The issue happens in new threads as well, not just old conversations\n\n\n\n\nExpected behavior:\n\n * Small tasks in a small repo should return much faster, especially in /fast mode\n\n\n\nAdditional note:\n\n * I checked the public status page, and it appeared operational when I looked, so this may be a model-specific or app-specific performance issue rather than a broad outage\n\n\n\nWould the team be able to check whether there is any current latency issue affecting GPT-5.4 in the Codex app on Windows?\n\nThanks.\n\n",
"title": "Codex app on Windows became extremely slow today even in /fast mode"
}