Replies dated this day in the collected sample: 100. Collected since 5 Oct 2026: 670.
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.
-
Instant steering is nice, but my Dot cloud PC wiped an Obsidian plugin three times yesterday: 17:31 source vault gone, 20:40 rolled back to an old one, 22:27 the restored vault and Obsidian vanished. When is that fixed?
@AShozak on X ↗ · 00:13 UTC
-
GitHub access is still broken. My sandbox can't even fetch the latest commit because of a proxy error. Others are reporting the same issue. Please fix the basics before announcing more upgrades.
@forevermirana on X ↗ · 00:15 UTC
- A problem
Didn't receive a bank reset
@moritziom on X ↗ · 00:46 UTC
- A problem
It actually refuses everything, it's actually insane something definetly broke.
@dyneteve on X ↗ · 00:48 UTC
-
Error: account/read failed during TUI bootstrap: account/read failed: workspace routing discovery timed out (code -32603)
@metalagman on X ↗ · 01:24 UTC
- Mixed
what's more... I really feel that compared to last month, the usage of pro5x&pro20x(Now pro10x) reduced a lot...
@FaustZ119 on X ↗ · 01:43 UTC
- An idea
Speed is great I guess, but the biggest issue for us is when we hit out limits we can no longer work. Should make GTP-6.1 Luna free in codex without limit so there is at least something reliable to fall back on when limits are hit for the big boys💪
@HoebenChris on X ↗ · 01:59 UTC
- A problem
Codex has recently been stopping on its own during the mission unless the 'goal' command used. Is anyone facing that issue ?
@DCENIAC on X ↗ · 02:01 UTC
- Mixed
not great tbh, fixed 1 thing, broke another
@diablovoxx on X ↗ · 02:16 UTC
- A problem
I've been experiencing frequent connection issues when trying to use my dot in Codex. It often fails to connect, even after multiple attempts, which really disrupts my workflow. Is this a known issue? Are there any plans to improve connection stability?
@kVUOLEzytNspt60 on X ↗ · 02:20 UTC
- A problem
I used up my 5-hour quota in just 45 minutes.!
@yyming487 on X ↗ · 02:23 UTC
- A problem
I'm having a desktop and browser opening loop when try to installing this on Mac. Press the install plugin on chatGPT app it opening browser and on browser it ask to open the app. Any other facing same issue?
@heyitskaiz on X ↗ · 02:26 UTC
- A win
4 update and a banked reset. Not bad.
@MP652300 on X ↗ · 02:29 UTC
-
Bruh y'all broke the Windows sandbox today. Can't get anything done 😭
@hypothesheesh on X ↗ · 02:37 UTC
- Mixed
Thanks for yesterday's chat model update, and for giving us a reset when I was already very happy with it. It's the reason I upgraded from Plus to Pro. Today's update, though, honestly doesn't mean much. Could we get a few more reset cards? That might even push me to…
@moxicoApple on X ↗ · 02:49 UTC
- A problem
Frequent message: selected model is at capacity. please try a different model. Please optimize this issue as soon as possible.
@jzhu82 on X ↗ · 02:51 UTC
- A problem
Why does my codex have the problem of forced exit after some versions are updated?
@dax7777777 on X ↗ · 02:55 UTC
- A problem
On the $ 100 plan, GPT-6.1 Sol Ultrafast is not available for me. Reset needed.
@realhanma on X ↗ · 03:03 UTC
- A problem
You told us "each day we’ll either ship one thing that is a clear improvement and relevant for most codex/work users or ship a full reset". This (GPT 6.1 Ultrafast) apply only to Pro 500. So not MOST CODEX USERS. Instant steering isn't a big deal as you believe. Reset.
@gxlmo on X ↗ · 03:06 UTC
- An idea
Hey @OpenAI, why can't ChatGPT access my previous chats even when I share my own conversation link? It asks me to copy everything manually. A user-authorized “Continue from another chat” feature would make long-term projects much easier.
@NithyonX on X ↗ · 03:52 UTC
- A problem
I think the issue that most needs improvement is that the quota isn't enough...Raise your hand if you agree
@mushengziyue on X ↗ · 03:55 UTC
- A problem
I'm experiencing issues with the Codex input field on macOS where selecting text with the mouse doesn't always work correctly.
@LoganLuoco on X ↗ · 04:01 UTC
- A problem
What the fuck is going on with Codex usage limits? My 20x usage reset yesterday. Just TWO sessions (Astra Medium + Astra XHigh) burned through my entire allowance in under 9 hours. $200/month for this? This is fucking ridiculous.
@0xjohnfx on X ↗ · 04:05 UTC
- A problem
I can not get my chat replacement to function. GPT-6.1 Sol is complete shit. They claim 5.6 is still available but impossible to lead.
@EncoreModern on X ↗ · 04:09 UTC
- A win
Eureka! Switching permissions from "Custom" to "Approve for me" in the Windows desktop app fixed it for me!
@maharris77 on X ↗ · 04:24 UTC
- A problem
Hmm not seeing my reset
@erikrogne on X ↗ · 05:10 UTC
- A problem
Okay, so as a $200 subscriber who doesn’t even have access to the ULTRAMOTHERFUCKINGFAST tier, I went back and checked regular GPT-6.1 Sol in Codex. The result? Still slow. The Day 1 speed improvements look like they got rolled back, and the whole Codex experience feels…
@yuhaijun on X ↗ · 05:51 UTC
- A problem
超级慢
@David217820 on X ↗ · 06:14 UTC
- Mixed
All ChatGPT subscribers are affectedThe Juice value is the length of the chain of thought that an AI model is allowed to use when running!When users use the API directly and enable Max Thinking on GPT-6 Astra,the Juice value is 832, while it’s only 28 in the Codex subscription.
@ranaz544127 on X ↗ · 06:25 UTC
- A problem
selected model is at capacity!!!!! PLEASE
@ZZZSSS00468268 on X ↗ · 07:02 UTC
- A problem
Actually, it's slower than before.
@sylvenPR on X ↗ · 07:15 UTC
- A problem
selected model is at capacity all day!
@FNnaclex on X ↗ · 07:17 UTC
- A problem
所以能不能再给一次重置,我的额度一下子就耗光了,重置卡用不到两个小时又用完了
@yuan98525981646 on X ↗ · 07:22 UTC
- A problem
我他妈感觉我的sol6.1跟死机了一样
@a12085424 on X ↗ · 07:23 UTC
- A problem
Your reset mechanism is an absolute disaster. Whether it's the delayed reset logic or the completely useless Plus reset crippled by the 5-hour limit, both sides are a mess. Until you actually fix these core restrictions, this feature will forever remain pure trash.
@hanakamiov on X ↗ · 08:22 UTC
- A problem
你能注意到我吗?这几天6.1sol 和 6astra 像个智障一样,什么问题都解决不了,浪费额度!
@ggdyi5 on X ↗ · 08:42 UTC
- A problem
Can someone show me where the banked reset is? Can’t find it
@ripeti_app on X ↗ · 08:53 UTC
-
Codex slow like a turtle always said reconnecting
@Rockbyway on X ↗ · 08:57 UTC
- A problem
I haven't seen the reset yet,why?
@tanner_hzmljx on X ↗ · 09:02 UTC
- A win
Solved with workaround, ask it to use git bash instead, and set it in settings. WSL might work too.
@Ingenitu on X ↗ · 09:04 UTC
- A problem
tibo Selected model is at capacity. Please try a different model. why!!!!!!!!!!!!!!!!!
@xueyue1787103 on X ↗ · 09:17 UTC
- Mixed
因为标准速度已经被降到了 6 tok/s,8x 差不多就是 49 tok/s 了
@Plasticine9750 on X ↗ · 09:51 UTC
- A problem
that is his definition of getting more work done. it's running for longer and drawing tokens slower. but in reality you're just waiting for shit. claude draws faster but actually gets your shit done. i have the same issue with codex right now. ive cancelled my 200 plan
@u83205 on X ↗ · 09:52 UTC
- A problem
even with 2, $500 dollar plans, can't make a full week without GPT 5.6 Sol Pro being throtted to medium. I'm done, this 2 year user is switching.... I reall held off considering the size of my project, but this is just ridiculors. $1000 a month and i stll can't work
@twatholio on X ↗ · 10:48 UTC
- A problem
Hi Tibo, I'm having major issues with using the cloud for a project. individual threads can't talk to each other which seems daft - it's difficult to set things up push completed work from one to the other and more challenging to have a local thread work with cloud threads even…
@DrFreeman666 on X ↗ · 10:50 UTC
- A problem
Despite the launch of dots, Intelligent UI still doesn’t work in dots chats. Also, voice conversations in dots aren’t as smooth as in Chat. Instead, they feel incomplete, just like in Work.
@mt_tsurezure on X ↗ · 11:55 UTC
- A problem
I haven’t received any why?
@SLKtom1 on X ↗ · 11:56 UTC
- A problem
btw did you guys quietly nerf limits on Sol 6.1? Today it feels like twice faster in spending my limit. No fast mode. Same reasoning as yesterday. Codex is clean, no plugins/MCPs working.
@jr_kurt70333 on X ↗ · 12:31 UTC
- A problem
Brother RIGHT NOW AS I TYPE THIS Sol 6.1 in "Fast" mode is STILL DOG SLOW. Can you signal to me you are aware of this Tweet and/or problem? It's the same BS every day now during the working day hours for GMT (8am-5pm) which is the middle of the morning for OpenAI staff. Explain?
@amihail01 on X ↗ · 13:05 UTC
- A problem
I didn't get one
@Aloufer1 on X ↗ · 13:16 UTC
- A problem
this only works once but also i don't want to babysit. i guess should add it to agents md but i prefer gaslighting the model instead hehe
@1000xcat on X ↗ · 13:58 UTC
- A problem
Getting an average of 23 tokens per second after your "speed boost" from day 1 of "improvements"
@SharkThePark on X ↗ · 14:07 UTC
- A problem
Codex is too slow
@hahagewhahagewh on X ↗ · 14:07 UTC
- Mixed
I’m getting ~150 tok/s in pi, what harness are you using?
@Chris_Falck on X ↗ · 14:22 UTC
- A problem
the regular folks didn't get anything yesterday, your promise was a reset every day we don't, there has been no rest.
@PromoteABook on X ↗ · 14:25 UTC
- A problem
Pretty bad, considering the only decent upgrade was increasing the speed of Sol. And obviously you made it slower not faster, so the one upgrade was just a lie. So in reality you've give out a banked reset and a forced reset that messed up a load of people's budgeting. 2/10.
@danny_______ on X ↗ · 14:26 UTC
- A problem
I'm a free-tier Codex user, watching these resets like everyone else 😂 One day I'll get to celebrate too!
@yura_oha on X ↗ · 14:26 UTC
- A problem
it sucks even more. takes me over 50 turns at each ui adjustment. have to spell out the obvious. for your delightedness built on my agony, the harness steering is a massive issue. it does not know how to adds on. it only change the direction drastically
@mhgilak on X ↗ · 14:43 UTC
- A win
They've fixed it, I was testing it on their Playground API (web version), now instead of 49 tok/s I'm getting 249 tok/s there. :)
@SimonasLTU1 on X ↗ · 14:47 UTC
- A problem
I hate to put it this way, but I don't notice any improvement
@Jess1316326 on X ↗ · 14:49 UTC
- A problem
Oh? I didn't any banked reset... I only got the reset from day 2🤯
@RanNi91685684 on X ↗ · 15:18 UTC
- Mixed
15 tokens per second rn for me
@James56009782 on X ↗ · 15:29 UTC
- A problem
Having to beg Tibo for resets just to make a model usable is ridiculous. I pay $100 each for ChatGPT and Claude, yet I can use Claude 24/7 without hitting limits. And even if I do, I get $250 in credits, $100 in API credits, plus a free banked reset.
@WilliamH0789 on X ↗ · 16:12 UTC
- Mixed
Currently running at 20TPS....
@illogicparadox on X ↗ · 16:50 UTC
- Mixed
50 TPS? Maybe only on your account, connected directly to the data center. 22-26 TPS
@diego_dion58017 on X ↗ · 17:00 UTC
- A problem
It's only (currently) showing on Work, on the web. Any tricks to get it working on codex desktop? On Version 26.1007.21159. I've tried to kick things off on web, move to desktop, switching from astra (ultrafast) to sol to see if it was a UI bug maybe; but no dice. Possibly a…
@asherhunt on X ↗ · 17:13 UTC
- A problem
Lol mine has been shitting the bed ever since I made this post. It is like it heard me j/k. Connections aren't working with it anymore. Basically stopped using it. Worked amazing on launch day was loving it.
@odin2128 on X ↗ · 17:26 UTC
- A problem
Thank you for the improvements, Tibo! As a Pro 500 subscriber, I’m seeing two issues with GPT-6.1 Sol ultrafast: the option sometimes disappears from the model picker, and when I can select it, it doesn’t feel faster. Is this known and being worked on, or am I missing something?
@MountainZeb on X ↗ · 18:15 UTC
- A problem
I want to try using it, I really do. But I can't get it to use in peace the creds I gave it in an env. If I gave it a GH PAT, I expect it to use the GH PAT; instead idk, it doesn't see it, it gets 403 from the OpenAI proxy anyway, autoreview blocks it or the moon was bright.🤪
@engelnyst on X ↗ · 18:21 UTC
- A problem
It still has the issue that every. single. time. you change a setting in the environment, you can't continue any of your chats. You guys shipped a computer that when you change a setting, you have to throw the entire thing away. Brilliant.
@voidprimer on X ↗ · 18:25 UTC
-
Atlassian appears to be removed now. Thank you.
@justynclark on X ↗ · 18:37 UTC
- A win
There has been an update that fixed this issue for me
@jamiefrost1279 on X ↗ · 18:56 UTC
- A problem
Didnt recived it yesterday, didnt recived it today. on 100$ plan, and tokens are burning way faster since last update =S
@gruzrh on X ↗ · 18:58 UTC
- An idea
with all the banked resets we need to be able to use more volume with our 5 hour limit :) currently its around 20% of weekly maybe make it 35%
@danison_1337 on X ↗ · 19:12 UTC
- A problem
No, it’s not. Create_thread tool is disabled and I can’t work without copy pasting stuff
@xAleAguilar on X ↗ · 19:35 UTC
- A problem
Except in chat, where you still can't steer it.
@s4m_bloch on X ↗ · 19:48 UTC
- A problem
I’ve been facing the issue of my codex app shutting down every 15 minutes. I’m on windows 11 and I’ve tried repairing the app, resetting it, uninstalling then reinstalling it. None of it has worked. @thsottiaux we need a fix
@theonlytobe1 on X ↗ · 19:48 UTC
- A problem
Lately, when I'm working in my browser, this message pops up and won't let me type or do anything.
@maix360 on X ↗ · 20:14 UTC
- An idea
It’s great to see OpenAI releasing new features almost every day. But at this point, the one update I really want to see is a truly unified ChatGPT experience. GPT-6 is impressive, but there’s still a fundamental user experience problem. I can spend hours working on a project…
@o_fab_o on X ↗ · 20:23 UTC
- A problem
gpt-6.1 sol in Codex is still painfully slow, barely 20+ TPS. How does adding dots to mobile meaningfully help Codex/Work users? If there's no real improvement today, can we at least get the full reset as promised?
@LowWeikee on X ↗ · 20:28 UTC
- A problem
Dots are so lazy I spend more time asking them why they stopped working and telling them to get back to work than it would take me to just do the tasks myself.
@octa2424 on X ↗ · 20:29 UTC
- A problem
Too bad Codex is still laggy and Astra can rewrite Codex in Rust for Windows . Astra is too dumb 💀
@Sovex666 on X ↗ · 20:38 UTC
- A problem
It was working on release in desktop but stopped working a few hours later.
@AlphaMFPEFM on X ↗ · 21:09 UTC
- A problem
wtf is up with the recent update?? the app on both of my machines is fucked. Dot responses dont show up until I follow up after thinking it doesnt respond, and my codex's wont connect to each other dav 5 is a wash. give us a banked reset plz
@EthanSwzr on X ↗ · 21:23 UTC
- A problem
which “Day X” do we get a working Chrome extension? 😭 app says install the extension, extension says install the app. reinstalled both, still cooked. judging by the reviews, it’s not just me
@dannleed on X ↗ · 21:30 UTC
- A problem
lol I still can’t even rename my dot it keeps saying error
@AOE234 on X ↗ · 21:35 UTC
- A problem
But I’ve already created my dot…. And my dot was already creating and working in my local codex threads (albeit it’s been buggy)…… What is new here other than bug fixes?
@Technop54777070 on X ↗ · 21:50 UTC
- A problem
I never got my banked reset on my Pro account , did anyone else ?
@wolfofdena on X ↗ · 22:02 UTC
- A problem
Give pro users full access to advanced models. Nerfed models that over think everything, then realize they are wrong. Say they are wrong. Then don’t fix what they just wasted usage on is the codex death spiral. Hard time justifying paying $500 a month for 5x more usage that gets…
@Sell_The_Rip on X ↗ · 22:04 UTC
- A problem
I’m on the $500 Pro plan. GPT-6.1 Sol’s Ultrafast (8×) option appeared yesterday but disappeared after a few hours of use. Why, and how can I get it back?
@xuyansheng on X ↗ · 22:16 UTC
- A problem
any chance you invest into quality of existing functionality, too? deep research is broken for some people since weeks (0 citations, 0 web search). Reporting issues is broken. See also
@_valoh_ on X ↗ · 22:47 UTC
- A problem
these are really good improvements, but we need a permanent solution for usage limits. I'm a pro max package user and still can't use my limits more than 1 day and I have to check when @thsottiaux will provide a codex reset. Can you guys focus on how to make limits efficient on…
@yerenkucuker on X ↗ · 22:53 UTC
- A problem
ill take the resets because nothing is fucking useful on my end. everything is fucking broken in chat and work the down time shows it and then its even worse the guard rails on cyber even with verification and tac regressed and support cant fucking help shame on you
@JamesGo47937709 on X ↗ · 22:53 UTC
- A problem
Why can't my Dot link with Codex on my local computer? I'm unable to use Dot to orchestrate the local Codex.
@AngelKittybing on X ↗ · 23:17 UTC
- An idea
You're the best, but we need the $200 tier because the $500 tier goes from 100% to 0% within a day, not enough for a week.
@bichenfan on X ↗ · 23:30 UTC
- A problem
Deleting a chat conversation on the website takes too long; hopefully, you'll fix that.
@isNhanh on X ↗ · 23:43 UTC
- A problem
Hey @keyanzhang @sharifshameem — when will Plus users get GPT-6 Luna and Sol in local Codex, like we already have in Work on the web? On Windows Codex Desktop, they still don't appear in the model picker, and gpt-6-luna also reports that it's unsupported with a ChatGPT account.…
@Mryan2005_1 on X ↗ · 23:48 UTC
- A problem
Recently Codex feels like an awkward mess — it’s like wearing grandma’s socks paired with high‑heels; it can barely move, slow and clunky.
@ZhaoqiMa64905 on X ↗ · 23:57 UTC
- Mixed
we really dont need this. vscode copilot had this ages ago. just give that reset card. obey the rules :) dont try to pull that on me
@AlbertSmitiz on X ↗ · 23:58 UTC
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.