{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreibyvhr4g4ppyltwfqz6nymfqyuec7zjbxdg4f7yfi7f6ncasejq6a",
"uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3modsolj7cab2"
},
"path": "/t/codex-rate-limits-discussion-thread/1378553?page=19#post_394",
"publishedAt": "2026-06-15T17:18:32.000Z",
"site": "https://community.openai.com",
"textContent": "Hi OpenAI / Codex team,\n\nI believe my Codex Pro quota is being metered abnormally. This looks like a quota accounting / bucket assignment / cached-token metering regression, not simply heavy usage.\n\nCurrent visible status from Codex /status:\n\n * 5h limit: 95% left, resets at 00:33 on June 16, 2026\n * Weekly limit: 18% left, resets at 22:16 on June 21, 2026\n * In other words, the weekly Codex quota appears to be ~82% used.\n\n\n\nThe reset timestamp in my usage snapshots matches this weekly window:\n\n * weekly reset: June 21, 2026 at 22:16 Europe/Moscow\n\n\n\nI inspected Codex session usage snapshots and found a mismatch between visible quota burn and logged token usage.\n\nMain session investigated:\n\n * Thread/session id: 019e7415-b691-7353-a830-d2231b1c5598\n * Thread name: Add branding prototype\n * Source: Codex Desktop / VS Code\n * Model: gpt-5.5\n * Plan type in usage snapshots: pro\n\n\n\nToken summary from this session:\n\n * token_count events: 434\n * cumulative total_tokens at last snapshot: 55,532,612\n * input_tokens: 55,337,946\n * cached_input_tokens: 53,562,752\n * output_tokens: 194,666\n * reasoning_output_tokens: 68,862\n * cached input is ~96.8% of input tokens\n\n\n\nBy date:\n\n * May 29, 2026: about 40.23M total_tokens delta\n * June 15, 2026: about 15.10M total_tokens delta in the parent thread\n\n\n\nOn June 15, this parent thread spawned three subagents:\n\n * 019ecb79-9c9a-7832-a90c-f39f698b115b / Ops\n * 019ecb79-b756-75d2-bd09-bd1555113a7f / Main\n * 019ecb79-d13d-7070-babf-1770c3fea07c / Vector\n\n\n\nThe subagent sessions appear to include copied cumulative parent context, so their final total_tokens should not be naively added as separate new usage. Estimated post-fork new token deltas:\n\n * Ops: no post-fork token_count delta visible after 13:30:30Z\n * Main: ~1.19M total_tokens\n * Vector: ~0.37M total_tokens\n\n\n\nSo for this investigated parent thread on June 15, the visible logged usage is roughly:\n\n * parent delta: ~15.10M total_tokens\n * active subagent deltas: ~1.56M total_tokens\n * total estimate: ~16.66M total_tokens\n\n\n\nI did not find a literal codex-auto-review source in these snapshots. The visible usage here appears to come from normal agent and subagent execution.\n\nI also inspected the current session that matches the visible /status quota state:\n\nCurrent session:\n\n * Session id: 019ecc3a-789f-7f51-98fc-773692c51cde\n * Model: gpt-5.5\n\n\n\nIn this current session, over a short diagnostic window:\n\n * first token snapshot: June 15, 2026 17:01:07 UTC\n * latest token snapshot: June 15, 2026 17:10:55 UTC\n * total_tokens delta: ~3,021,197\n * input_tokens delta: ~2,997,763\n * cached_input_tokens delta: ~2,716,416\n * output_tokens delta: 23,434\n * reasoning_output_tokens delta: 8,232\n\n\n\nDuring that same usage window:\n\n * 5h usage moved from 3% used to 7% used\n * weekly usage moved from 81% used to 82% used\n\n\n\nThis is the core concern: a relatively small logged delta of about 3M total_tokens moved the weekly quota by about 1 percentage point, while the account is already showing ~82% weekly usage. If that ratio is representative, it implies an effective weekly allowance of only around ~300M total_tokens, which seems inconsistent with a Pro account and with prior Codex usage expectations.\n\nPlease investigate the backend quota accounting for this account, especially:\n\n 1. Whether cached input tokens are being charged against included weekly quota at an abnormal or full weight.\n 2. Whether long-running sessions, compaction, or copied parent context in subagents are being counted repeatedly.\n 3. Whether subagent/forked sessions are double-counting inherited cumulative context.\n 4. Whether the account is assigned to the correct Pro quota bucket.\n 5. Whether GPT-5.5 local usage is being metered with the intended included-limit weighting.\n 6. Whether the UI is mixing used vs remaining percentages or stale snapshots.\n 7. Whether there is a recent quota/metering regression around the current weekly reset window.\n\n\n\nWhat I need:\n\n * Account-level review by the Codex quota/accounting team.\n * Explanation of why the weekly quota is already ~82% used.\n * Restoration or adjustment of the weekly quota if this was incorrectly metered.\n * Better usage breakdown in the dashboard: interactive agent vs subagent vs auto-review/background usage, and new usage vs inherited/cached context.\n\n\n\nI can provide screenshots of /status and additional aggregate usage numbers if needed.\n\nAlso ccusage show this data - 80% of weekly limits on 20x sub lost in 1 day with 500m tokens\n│ 2026-06-15 │ - gpt-5.5 │ 26,512,423 │ 1,423,718 │ 430,733 │ 539,149,312 │ 567,085,453 │ $444.85 │",
"title": "Codex Rate Limits Discussion Thread"
}