I’ve been consistently using the OpenAI official macOS app with ChatGPT (o1 Pro) for months, regularly pasting large content ranging between 100k–130k tokens without any issue. This workflow was extremely beneficial for my projects.
However, just a few minutes ago, the macOS app prompted me to update, and after completing this update, I’m encountering a severe paste size limitation. When attempting to paste content around my usual size (100k-130k tokens), the app now immediately gives me an error: “Paste size limit exceeded. The content you are trying to paste is too long.”
To troubleshoot, I progressively reduced my paste size:
At 80k tokens: Same error.
At 45k tokens: Still the same error.
At 29k tokens: Finally able to paste successfully.
This new limitation is significantly impacting usability, especially considering that pasting the original 130k tokens directly into o1 Pro via the web interface still works.
This indicates that the macOS app specifically has introduced a far more restrictive paste size limit.
Could anyone from the OpenAI team clarify why this significant restriction was introduced in the latest update? If the underlying model supports content above 100k tokens, why impose such a restrictive limitation in the app?
I strongly recommend reconsidering this limit or providing an option to adjust it, as it’s seriously hindering workflow and productivity for power users.
Seconded! Severe degradation for pro users, which is a bummer given that the MacOS app is quite nice these days.
I like the idea of a preemptive token counter to catch oversized messages before submission-- this would generally be more elegant than the previous behavior; where the interface would reject the message only after submission. A paste limit is a fine version of this (albeit of course the actual token limit is the whole of the message, not just the pasted content).
But the current version is just incorrect, reflecting the lower-tier-plan input limits and not the Pro limits. Very annoying!
Pro user here. I’m paying $200/month for an entirely arbitrary 128,000 byte paste limit on macOS that is not enforced on iOS or the web. I can paste multiple 127,999 byte chunks into the input field, which makes me think this must be a bug: any sincere attempt to limit the load on OAI infra would presumably have imposed a character limit.
This is really frustrating, and OAI’s macOS release notes page is a mess: there are no notes for version 1.2025.077 (1742622997) but the only reason I know that is because I spent an hour figuring it out. The release notes page has no version data, it just lists the date of release and any major features.
I know it’s kinda passé, but here’s what 4o thought of the page when I asked it “Is this page well-designed and helpful?”
Recommendations
Add a TOC or version index at the top .
Categorize entries with tags or icons: Feature, Fix, UI, Security, etc.
Link versions to source/update logs, especially for developers or power users.
Include visuals or GIFs for major UX/UI changes (especially on Mac apps).
Enable in-page search/filter or collapse older versions for readability.
Verdict
Is it well-designed?Functionally, yes — aesthetically and structurally, no.
Same issue here — this new paste limitation in the Mac desktop app is a big step backward. I regularly use ChatGPT to analyze large blocks of code and documentation, and this change breaks that workflow.
Really hope this gets reverted or made configurable.
Slightly different problem but with the same Version 1.2025.077 (1742622997) - I updated it earlier today and it’s crashed three times now with no warning and no apparent reason (each time when I was in the middle of typing a prompt). Anyone any ideas why? I’m using MacOS Sequoia 15.3.2. Thanks
Same here, been using ChatGPT MacOS desktop app as a daily basis, paying 234€ / month as it was so handy to copy long text, now, without further notice the limitation is applied. Is this a deliberate movement? will it work like this from now on? Any information on this?
Just want to get in on this thread as I too am experiencing the Mac app paste size limit restrictions. This has severely hindered by ability to use this app. I am hoping that this is an error that will get fixed soon.
I am now doing most of my more complicated stuff with Claude. I am a 200$/month user. I can’t talk with o1-pro about a 4.000 line code segment anymore. I was using 80% openAI and 20% Claude. Now it is the other way around.
I am very disappointed and also think this is openAI showing “We really don’t care about our customers. We restrict the capabilities of our paying users because we need the computing power for a new product. If this breaks old customers workflow, we don’t care.”
Im curious how aware of the problem they even are. Theres been a minor update since this issue started happening. Either they dont read this forum or its intentional.
Edit: unsubbed from pro, back on plus and now use another company for the long context window usecase
Recommendations