Codex 28 · Day record · 6 Oct 2026

Day 2 of 28

Closed at 00:00 UTC on 7 Oct 2026 Counted only from activity timestamped inside this UTC day. Later votes, answers and experiences do not change it; only an entry in the corrections log can.

Revised 25 Oct 2026, 07:25 UTC after an entry in the corrections log. Corrections log →

Download card

The day’s change

Recorded update

Improvement

Day 2.1: Auto-review is now free with ChatGPT sign-in

Day 2.1: Tibo announces that Auto-review is now free for all users signed in through a ChatGPT account, and that its review usage no longer counts against their plan. Enable it in Settings > General > Permissions > Auto-review, as shown in Tibo’s follow-up screenshot.

Read the whole note

A separate agent reviews actions that request permission to cross the sandbox boundary, checking for high-risk actions and actions outside the user’s intent. The change announced today is free access and no plan usage for these reviews; Auto-review itself was already available.

OpenAI’s technical explanation ↗ Try it on your own tasks and vote on whether this change is useful.

In a follow-up to the Auto-review announcement, Tibo recommends using Auto-review instead of full access and calls it “not a trade-off anymore”. This is his recommendation, not Rantr’s independent security assessment. Source ↗

Tibo later clarifies that Auto-review is the “Approve for me” permission mode: open the permissions menu below the composer and select “Approve for me”. This adds his later UI path to the settings path in the announcement; Rantr has not independently checked it in every app version. Source ↗

In the Day 2 roundup, Tibo calls Auto-review “Approve for me”, says it is included without using plan allowance, and says it could previously account for 2–10% of a plan when used. That range is Tibo’s claim, not a measurement by Rantr. Source ↗

Source ↗

Fewer than 10 votes by the close of the day; the tally is in the open data, not shown here.

Improvement

Day 2.2: Easier qualification for higher OpenAI API rate limits

Tibo marks this as Day 2.2, a second announcement on 6 October. OpenAI says its five paid API usage tiers are becoming three: Build, Launch and Grow. The total API payments threshold to qualify for Grow, the new highest tier, is $500, down from $1,000 for the previous highest tier.

Read the whole note

OpenAI says organizations already on a paid tier move to their new tier automatically, with no action needed. Check your organization’s Limits page for its actual qualification requirements and model-specific limits.

This announcement concerns API tier qualification. It does not announce a ChatGPT/Codex subscription reset, lower token prices or a doubling of usage. Judge whether it helps your own API workflow; Auto-review remains a separate Day 2.1 entry with its own votes.

Official announcement ↗ Automatic migration follow-up ↗

Source ↗

Fewer than 10 votes by the close of the day; the tally is in the open data, not shown here.

Improvement

Day 2.3: Meetings plugin for notes and follow-up

Tibo announced meeting notes as Day 2.3. The official ChatGPT announcement says the Meetings plugin creates personalized meeting summaries and next steps in ChatGPT Space, with options to keep notes private or share them and ask ChatGPT for follow-up work. Available in beta for Pro and Business users in the ChatGPT desktop app on macOS; Enterprise is described as coming soon. This is a meeting workflow feature, not a usage reset. Official announcement ↗

Source ↗

Recorded after the day closed (7 Oct 2026, 03:10 UTC), so no votes fall inside the day.

Fewer than 10 votes by the close of the day; the tally is in the open data, not shown here.

Improvement

Day 2.4: Decisions API enters public beta

Tibo announced the Decisions API as Day 2.4, quoting OpenAI Developers: developers can use it to choose a model, tool, or action in near real time. OpenAI describes it as available to all developers in public beta and claims decisions up to 10x faster than GPT-6 Luna through the Responses API; this is a vendor comparison, not an independent benchmark. Tibo says they will use it to improve the app experience. That app integration is a future intention, not a confirmed shipped app feature or usage reset.

Read the whole note

Replying to a question about whether the Decisions API accepts images, Tibo answers “It does”. This confirms his statement about image input for that API; it does not announce a new Codex model or another daily release. Rantr has not independently tested image input. Source ↗

Source ↗

Recorded after the day closed (7 Oct 2026, 03:10 UTC), so no votes fall inside the day.

Fewer than 10 votes by the close of the day; the tally is in the open data, not shown here.

Reset

Day 2: Tibo says the reset has been processed after the community poll

Tibo announced that the reset has been processed following the Day 2 vote. In a follow-up he says the four improvements remain and refers to the next releases as Day 3; this reset is therefore recorded as the Day 2 outcome, although the announcement was posted on 7 October UTC.

Read the whole note

The completed Day 2 X poll shows 74,565 votes: 76% chose needs a reset and 24% chose good day. A separate calibration poll shows 28,471 votes: 80.1% chose needs a reset and 19.9% chose great release. These are self-selected X poll results, not Rantr votes or a representative survey of all Codex users.

The reset is recorded from Tibo’s announcement; Rantr has not independently verified its application to every account. Check your own usage balance and next reset time, then share what changed.

Day 2 poll ↗ Calibration poll ↗ Improvements remain / Day 3 follow-up ↗

Source ↗

Recorded after the day closed (7 Oct 2026, 04:23 UTC), so no votes fall inside the day.

Fewer than 10 votes by the close of the day; the tally is in the open data, not shown here.

Running totals

What changed since day 1

Totals since 5 Oct 2026, at the close of each day; today’s are so far. A total never goes down when a sample is recounted.

Replies on X

Replies to Tibo’s post published this day

A sample collected from X, not every reply. Each link opens the reply on X.

Replies dated this day in the collected sample: 128. Collected since 5 Oct 2026: 162.

X replies are found by search and each one is checked against its public post on X. Last read: 26 Oct 2026, 23:46 UTC. Not every reply is read.

1 X reply from this day is not quoted: Rantr does not have its words or held them back after automated screening. It stays on X.

Nothing this day for: OpenAI release notes, published experiences, corrections.

Method

How this record is made

Each day runs from 00:00 to 24:00 UTC. The record counts votes, release check answers, experiences, “Same here” and replies timestamped inside the day. When the day closes the record is frozen; it is recomputed only after an entry in the public corrections log, or when the day’s update is recorded late. Rantr is a neutral record: it does not take a side.

Corrections log → Open data and method →