{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreifndj7utpk7ddbwouxdsreasbzinlwu7onecwlyr7ci3k74ezagai",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mfworclyry22"
  },
  "path": "/t/why-are-completion-tokens-so-high/1375358#post_3",
  "publishedAt": "2026-02-28T16:50:24.000Z",
  "site": "https://community.openai.com",
  "tags": [
    "@_j"
  ],
  "textContent": "Why use reasoning at all for this? It sounds like a classical categorisation task that you could evaluate with a basic LLM without any internal “loops”.\n\nI second @_j suggestion of using 4.1",
  "title": "Why are completion tokens so high?"
}