Posting this because the current experience across ChatGPT and Codex has become unnecessarily confusing, and a lot of that confusion comes from inconsistent wording, inconsistent UI, and documentation that does not reliably match what users are actually seeing.
The main issue is that users are being forced to reverse engineer product behavior instead of being able to trust the docs and the UI.
A few concrete examples:
In ChatGPT on iPad, I was trying to find the old thinking control that used to let me switch between standard and extended thinking. It was gone from the normal place. The documentation and explanations around the newer model picker suggested there should be a Configure option or some other clear place to manage thinking effort. But in the actual app UI on my account, there was no Configure button at all. The visible options were Auto, Instant, Thinking, Pro, and Legacy models. So from a user perspective, the docs implied one thing while the app showed something else.
That creates a really frustrating experience because it makes users think they are missing something, using the app wrong, or overlooking an obvious control, when the real problem is that the communication is unclear or the rollout is inconsistent.
A second example is Codex model availability. The April 7, 2026 update said that for users signing in with ChatGPT, the model picker would no longer show gpt-5.2-codex, gpt-5.1-codex-mini, gpt-5.1-codex-max, gpt-5.1-codex, gpt-5.1, or gpt-5, and that those models would be removed from Codex for ChatGPT sign-in on April 14. But in the actual Codex app UI, those older models appeared back in the picker. So again, the announcement said one thing and the product surface showed another.
Another problem is how plan and model wording is presented. The naming across GPT, Codex, Instant, Thinking, Pro, Auto, legacy models, and the various paid plans is not clear enough. The language often sounds definitive, but in practice it seems to depend on platform, account state, app build, rollout bucket, and plan. Users are left guessing whether a missing feature is intentional, moved, renamed, hidden behind rollout logic, or temporarily inconsistent with the docs.
Even the phrasing around promotions and usage limits is hard to parse. For example, statements like “continues on its current Codex promotion through May 31” are too vague on their own and require users to cross-reference multiple pages just to figure out what that actually means in practice.
The bigger issue is trust. When documentation, release notes, and the live UI do not match, users stop trusting all three. At that point, even simple questions become hard to answer:
- Did a feature move?
- Was it removed?
- Is it account-gated?
- Is it plan-gated?
- Is the doc outdated?
- Is the UI in an experiment?
- Is the rollout partial?
- Is the wording just imprecise?
Users should not have to do that kind of detective work just to understand the products they are paying for.
What would help:
- Make docs explicitly state when behavior may differ by rollout, platform, plan, or account state
- Use exact UI labels that match what users will literally see on screen
- Include screenshots for major picker and settings changes
- Clearly distinguish between ChatGPT UI behavior and Codex UI behavior
- Stop using vague phrasing when describing availability, limits, promotions, or removals
- When docs are describing an intended future state rather than the currently visible state, say that directly
Right now the experience feels like the products are changing faster than they are being explained, and users are left absorbing the confusion. That is especially frustrating when the issue is not user error, but communication quality.
Please tighten the wording, keep docs synchronized with live UI, and be much clearer about what is changing, where, for whom, and when.
Thanks.