Skip to content

Codex Cloud Rate Limit issue #6008

Closed as duplicate
Closed as duplicate
Codex Cloud Rate Limit issue#6008
Labels
bugSomething isn't workingcodex-webIssues related to Codex Webrate-limitsIssues related to rate limits, quotas, and token usage reporting

Description

@lunamidori5
lunamidori5
opened on Oct 31, 2025

What version of Codex is running?

Cloud

What subscription do you have?

Pro / Teams

Which model were you using?

Cloud

What platform is your computer?

Arch Linux

What issue are you seeing?

1 task is maxing out the 5 hour rate limit and using 30% of my weekly limit

What steps can reproduce the bug?

Send a task to codex cloud

What is the expected behavior?

Being able to do my normal 10+ tasks a hour

Additional information

Here are some photos where you can see today is a showing odd usage...

Image Image

Activity

  1. lunamidori5
    added
    bugSomething isn't working
    on Oct 31, 2025
  2. github-actions
    added
    codex-webIssues related to Codex Web
    on Oct 31, 2025

  3. github-actions commented on Oct 31, 2025

    @github-actions
    github-actionsBot
    on Oct 31, 2025 – with GitHub Actions
    Contributor

    Potential duplicates detected. Please review them and close your issue if it is a duplicate.

    Powered by Codex Action

  4. etraut-openai
    added
    rate-limitsIssues related to rate limits, quotas, and token usage reporting
    on Oct 31, 2025

  5. Abenezer0001 commented on Nov 1, 2025

    @Abenezer0001
    Abenezer0001
    on Nov 1, 2025

    where can i see my usage like that

  6. techjoec commented on Nov 2, 2025

    @techjoec
    techjoec
    on Nov 2, 2025

    @tibo-openai I saw your reddit posts and such, but something is still wonky it seems. I did ~30 cloud tasks and the following CLI work (2 sessions) and hit my 5-hour on Pro? :(

    l1 is Primary, l2 is secondary. And the random secondary of 23% makes me curious too....Really scratching my head and very saddened.

    as of 4.30am UTC I have 1.5 hours left until my 5-hour limit resets. So I guess it reset around row #2? I probably had less cloud tasks run than my quoted 30 in that 5-hour too.

    Timestamp InputTokens CachedTokens Output Reason Tokens l1 l2 Cache% Out/In% Reason%
    2025-11-02T00:03:09.327Z 5,778,653 5,171,584 75,299 27,392 5,853,952 1.0 11.0 89.5% 1.30% 36.4%
    2025-11-02T00:10:00.675Z 296,827 242,176 11,240 8,384 308,067 2.0 12.0 81.6% 3.79% 74.6%
    2025-11-02T01:41:52.464Z 26,352,640 24,274,688 140,561 56,640 26,493,201 19.0 6.0 92.1% 0.53% 40.3%
    2025-11-02T02:20:49.620Z 3,057,121 2,729,472 17,211 10,240 3,074,332 33.0 10.0 89.3% 0.56% 59.5%
    2025-11-02T03:30:54.455Z 28,498,125 27,207,168 230,372 129,216 28,728,497 78.0 23.0 95.5% 0.81% 56.1%
    2025-11-02T03:48:55.153Z 9,546 0 71 64 9,617 95.0 29.0 0.0% 0.74% 90.1%
    2025-11-02T04:10:04.832Z — — — — — 100.0 30.0 — — —

  7. techjoec commented on Nov 2, 2025

    @techjoec
    techjoec
    on Nov 2, 2025

    where can i see my usage like that

    https://chatgpt.com/codex/settings/usage

    Change the drop down to Weekly

  8. techjoec commented on Nov 2, 2025

    @techjoec
    techjoec
    on Nov 2, 2025

    I just burned my next 5-hour usage in about 60 mins..These are 'end of session' stats just for a glance.

    Timestamp InputTokens CachedTokens Output Reason Tokens l1 l2 Cache% Out/In% Reason%
    2025-11-02T07:59:43.173Z 27,009,104 26,430,080 101,882 32,768 27,110,986 66.0 50.0 97.9% 0.38% 32.1%
    2025-11-02T08:07:03.551Z 3,682,497 3,186,048 36,323 20,160 3,718,820 76.0 53.0 86.5% 0.99% 55.5%
    2025-11-02T08:32:49.420Z 4,154,169 3,848,960 18,121 9,920 4,172,290 96.0 59.0 92.7% 0.44% 54.7%
    2025-11-02T08:36:26.810Z 66,570,198 62,489,088 238,626 140,352 66,808,824 96.0 59.0 93.9% 0.36% 58.8%
    2025-11-02T08:43:55.000Z 1,349,797 1,174,400 10,502 6,720 1,360,299 100.0 60.0 87.0% 0.78% 64.0%

    And about two dozen cloud messages. Again this is on Pro too. :(

  9. lunamidori5 commented on Nov 2, 2025

    @lunamidori5
    lunamidori5
    on Nov 2, 2025
    Author

    Update... goodness... two small tasks and it uses all of my usage?

    Image

    adding spacing
    Image

  10. techjoec commented on Nov 2, 2025

    @techjoec
    techjoec
    on Nov 2, 2025

    I am on .49 still, as I really fear the updates with a lot of these issues for the record. But I also see duplicating token_count entries in rollout like this constantly (some are duplicate token counts...about 10% in a sample of a few hours??):

    {
    "timestamp": "2025-11-02T00:13:05.946Z",
    "type": "turn_context",
    "payload": {
    "cwd": "/github/repos/XXXYYYZZZ",
    "approval_policy": "never",
    "sandbox_policy": {
    "mode": "danger-full-access"
    },
    "model": "gpt-5",
    "effort": "high",
    "summary": "auto"
    }
    }
    {
    "timestamp": "2025-11-02T00:13:06.253Z",
    "type": "event_msg",
    "payload": {
    "type": "token_count",
    "info": {
    "total_token_usage": {
    "input_tokens": 272648,
    "cached_input_tokens": 250496,
    "output_tokens": 1864,
    "reasoning_output_tokens": 1280,
    "total_tokens": 274512
    },
    "last_token_usage": {
    "input_tokens": 22482,
    "cached_input_tokens": 22272,
    "output_tokens": 43,
    "reasoning_output_tokens": 0,
    "total_tokens": 22525
    },
    "model_context_window": 258400
    },
    "rate_limits": {
    "primary": {
    "used_percent": 2.0,
    "window_minutes": 300,
    "resets_at": 1762059736
    },
    "secondary": {
    "used_percent": 12.0,
    "window_minutes": 10080,
    "resets_at": 1762471693
    }
    }
    }
    }
    {
    "timestamp": "2025-11-02T00:13:12.239Z",
    "type": "event_msg",
    "payload": {
    "type": "token_count",
    "info": {
    "total_token_usage": {
    "input_tokens": 295233,
    "cached_input_tokens": 272896,
    "output_tokens": 2215,
    "reasoning_output_tokens": 1472,
    "total_tokens": 297448
    },
    "last_token_usage": {
    "input_tokens": 22585,
    "cached_input_tokens": 22400,
    "output_tokens": 351,
    "reasoning_output_tokens": 192,
    "total_tokens": 22936
    },
    "model_context_window": 258400
    },
    "rate_limits": {
    "primary": {
    "used_percent": 2.0,
    "window_minutes": 300,
    "resets_at": 1762059736
    },
    "secondary": {
    "used_percent": 12.0,
    "window_minutes": 10080,
    "resets_at": 1762471693
    }
    }
    }
    }

  11. techjoec commented on Nov 2, 2025

    @techjoec
    techjoec
    on Nov 2, 2025

    Hit my weekly in like three or four five hour limits. Locked out til my pro is supposed to renew in six days. Guess that makes my decision to downgrade from pro easier. Considering a charge back with the bank given the mid term rug pull.

  12. techjoec commented on Nov 3, 2025

    @techjoec
    techjoec
    on Nov 3, 2025

    I just got home and sat down looking at the numbers...This is really stupid. I am under 24 hours into my week and hosed. And again, mucking with this stuff coming into a weekend..I love you guys pouring your heart and soul into this stuff, and I'm sure it wasn't your call to push to prod on a Friday, but W T F. And it's definitely broke as Web keeps flirting with me being at my quota, and near my quota, randomly. Hopefully GPT isn't doing the math on the backend here, as we know 3+7=banana randomly... 👎

    Per day breakdown per Codex usage report JSON (which doesnt even add up to 100% ??? )

    {
    "data": [
    {
    "date": "2025-11-01",
    "product_surface_usage_values": {
    "cli": 4.734244660872603,
    "vscode": 0.0,
    "web": 19.14151781164927,
    "slack": 0.0,
    "sdk": 0.0,
    "exec": 0.0,
    "github_turn": 0.06114338122898196,
    "github_code_review": 0.7460948304726966
    }
    },
    {
    "date": "2025-11-02",
    "product_surface_usage_values": {
    "cli": 5.380617548150413,
    "vscode": 0.0,
    "web": 39.7206329795752,
    "slack": 0.0,
    "sdk": 0.0,
    "exec": 0.0,
    "github_turn": 0.0,
    "github_code_review": 0.0
    }
    }
    ],
    "units": "percent"
    }

    And via the 'usage' JSON:

    {
    "plan_type": "pro",
    "rate_limit": {
    "allowed": true,
    "limit_reached": false,
    "primary_window": {
    "used_percent": 0,
    "limit_window_seconds": 18000,
    "reset_after_seconds": 18000,
    "reset_at": 1762147153
    },
    "secondary_window": {
    "used_percent": 100,
    "limit_window_seconds": 604800,
    "reset_after_seconds": 521436,
    "reset_at": 1762650589
    }
    },
    "credits": {
    "unlimited": false,
    "balance": "0",
    "approx_local_messages": [
    0,
    0
    ],
    "approx_cloud_messages": [
    0,
    0
    ]
    }
    }

  13. techjoec commented on Nov 3, 2025

    @techjoec
    techjoec
    on Nov 3, 2025

    And just like that, now I have an extra two hours added to my "week":

    About 90 mins ago:

    Resets Nov 8, 2025 7:09 PM

    Now:

    Resets Nov 8, 2025 9:02 PM

    It didn't even change to an even 2 hour jump....What the heck?

    Edit: Now it's back to 7:09:50 ish, it seems to hop around from 1-3 seconds of that so far now that I'm watching it via script.

  14. techjoec commented on Nov 4, 2025

    @techjoec
    techjoec
    on Nov 4, 2025

    So radio silence all day after all those google docs and reddit posts and all.. Really wish @tibo-openai or Embiricos would share something. Even an ack would be nice?? I saw this https://community.openai.com/t/issue-with-gpt-5-codex-limits-despite-having-pro-subscription/1359497/62 posted just before 7am Eastern today (Monday) but nada. Just tons and tons of folks with the same questions/concerns/complaints.

  15. etraut-openai commented on Nov 4, 2025

    @etraut-openai
    etraut-openai
    on Nov 4, 2025
    Contributor

    @techjoec, check out the update in this issue.

  16. etraut-openai commented on Nov 6, 2025

    @etraut-openai
    etraut-openai
    on Nov 6, 2025
    Contributor

    Thanks for your post. We're providing updates related to this issue in this thread. I'm going to close this issue to reduce duplication.

  17. etraut-openai
    closed this as duplicateon Nov 6, 2025

  18. lunamidori5 commented on Nov 11, 2025

    @lunamidori5
    lunamidori5
    on Nov 11, 2025
    Author

    @etraut-openai thank you all for your hard work

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't workingcodex-webIssues related to Codex Webrate-limitsIssues related to rate limits, quotas, and token usage reporting

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions