{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreiflzqaem2tzz3eion722u4souqy437hqoco7di4rblf74itetdkbe",
"uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mmljs3epn422"
},
"path": "/t/recent-performance-degradations-in-mobile-codex-gpt-5-5/1381598#post_2",
"publishedAt": "2026-05-24T07:37:59.000Z",
"site": "https://community.openai.com",
"textContent": "Yes 100%. GPT 5.5 doesn’t even think through anything properly before answering. But that’s not new for the latest model - it HAS gotten worse with time, however. And I’m using Pro and Thinking Heavy!\nBut codex? definitely makes a lot more mistakes lately",
"title": "Recent performance degradations in mobile Codex/GPT 5.5"
}