# ![@thsottiaux Avatar](https://lunarcrush.com/gi/w:26/cr:twitter::1953337039510003712.png) @thsottiaux Tibo

Tibo posts on X about if you, we are, open ai, model the most. They currently have [------] followers and [---] posts still getting attention that total [-------] engagements in the last [--] hours.

### Engagements: [-------] [#](/creator/twitter::1953337039510003712/interactions)
![Engagements Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1953337039510003712/c:line/m:interactions.svg)

- [--] Week [---------] -32%
- [--] Month [---------] +26%
- [--] Months [----------] +7,218%

### Mentions: [--] [#](/creator/twitter::1953337039510003712/posts_active)
![Mentions Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1953337039510003712/c:line/m:posts_active.svg)

- [--] Week [--] +2%
- [--] Month [--] -1.20%
- [--] Months [---] +7,050%

### Followers: [------] [#](/creator/twitter::1953337039510003712/followers)
![Followers Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1953337039510003712/c:line/m:followers.svg)

- [--] Week [------] +7.90%
- [--] Month [------] +37%
- [--] Months [------] +3,992%

### CreatorRank: [------] [#](/creator/twitter::1953337039510003712/influencer_rank)
![CreatorRank Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1953337039510003712/c:line/m:influencer_rank.svg)

### Social Influence

**Social category influence**
[technology brands](/list/technology-brands)  [finance](/list/finance)  [stocks](/list/stocks)  [social networks](/list/social-networks)  [exchanges](/list/exchanges)  [celebrities](/list/celebrities)  [countries](/list/countries) 

**Social topic influence**
[if you](/topic/if-you), [we are](/topic/we-are), [open ai](/topic/open-ai) #565, [model](/topic/model), [the first](/topic/the-first), [to the](/topic/to-the), [codex](/topic/codex) #57, [update](/topic/update), [bit](/topic/bit), [command](/topic/command)

**Top assets mentioned**
[Coinbase Global Inc. (COIN)](/topic/coinbase)
### Top Social Posts
Top posts by engagements in the last [--] hours

"Codex CLI now works with your ChatGPT login with generous GPT-5 use included in the plus and pro plans. $ brew install codex $ codex It's that simple"  
[X Link](https://x.com/anyuser/status/1957133984657481956)  2025-08-17T17:34Z 29.9K followers, 312.1K engagements


"@carmelyne Could you DM me a screenshot or any further info to help repro We'll have a look"  
[X Link](https://x.com/thsottiaux/status/1957297964600435094)  2025-08-18T04:26Z [----] followers, [---] engagements


"@wowwowza That info is out of date and was for the api keys. Lots more than that directly included in your ChatGPT subscription just codex login"  
[X Link](https://x.com/thsottiaux/status/1957298360442126599)  2025-08-18T04:27Z [----] followers, [---] engagements


"@kopertop @aniketmaurya That's unexpected which version are you on codex --version Should be $ codex --version codex-cli 0.22.0"  
[X Link](https://x.com/thsottiaux/status/1957484695580733709)  2025-08-18T16:48Z [----] followers, [---] engagements


"@juniorro16 @kopertop @aniketmaurya Same way you paste everywhere else Ctrl/Cmd+C then Ctrl/Cmd+V"  
[X Link](https://x.com/thsottiaux/status/1957913889078726715)  2025-08-19T21:13Z [----] followers, [--] engagements


"While we're working on our documentation here is how to use gpt-5 high codex -c model_reasoning_effort=high Super useful when you need that extra oomph of intelligence when your other model friend gets yourself into a pickle 👀"  
[X Link](https://x.com/anyuser/status/1958001126521073875)  2025-08-20T03:00Z 29.9K followers, 46.5K engagements


"Codex in ChatGPT now supports image inputs also coming shortly to the Codex CLI. Try it at Love using this to go from diagrams or mockups to implementation in one shot https://chatgpt.com/codex https://chatgpt.com/codex"  
[X Link](https://x.com/anyuser/status/1958744816784654500)  2025-08-22T04:15Z 29.9K followers, 45.6K engagements


"@Angaisb_ was just merged will be part of next release https://github.com/openai/codex/pull/2544 https://github.com/openai/codex/pull/2544"  
[X Link](https://x.com/thsottiaux/status/1958991592217616835)  2025-08-22T20:36Z [----] followers, [----] engagements


"Never worked with a more talented team feeling grateful and excited for the week ahead. It's codex time"  
[X Link](https://x.com/anyuser/status/1959358983120269727)  2025-08-23T20:55Z 29.9K followers, 48.8K engagements


"@rheza Send me the error what do you see"  
[X Link](https://x.com/thsottiaux/status/1959628518704472096)  2025-08-24T14:47Z [----] followers, [--] engagements


"Strong foundations compound. crazy how fast the tables switched from claude code to codex in my experience both are very good but gpt-5 just being better makes me lean towards codex crazy how fast the tables switched from claude code to codex in my experience both are very good but gpt-5 just being better makes me lean towards codex"  
[X Link](https://x.com/thsottiaux/status/1960105186619519375)  2025-08-25T22:21Z [----] followers, [----] engagements


"Image inputs landed in codex cli. Update to [----] to try it along with many other improvements"  
[X Link](https://x.com/anyuser/status/1960579534257820024)  2025-08-27T05:46Z 29.9K followers, 176.8K engagements


"@CarryFaze What would you like next"  
[X Link](https://x.com/thsottiaux/status/1960595381064016120)  2025-08-27T06:48Z [----] followers, [---] engagements


"@respectorr69 Yes we fixed a regression in how rate limits were calculated that was accidentally rolled out last Friday. It now works as intended. More info on exact limits shortly"  
[X Link](https://x.com/thsottiaux/status/1960705663731986635)  2025-08-27T14:07Z [----] followers, [---] engagements


"@mark_k We plan to ship daily"  
[X Link](https://x.com/anyuser/status/1960776555036389509)  2025-08-27T18:48Z 29.9K followers, 22.5K engagements


"👀 https://marketplace.visualstudio.com/itemsitemName=openai.chatgpt&ssr=false#review-details https://marketplace.visualstudio.com/itemsitemName=openai.chatgpt&ssr=false#review-details"  
[X Link](https://x.com/anyuser/status/1960803511651983455)  2025-08-27T20:36Z 29.9K followers, 31.5K engagements


"@cogito_matt @mark_k @OpenAI codex together with codex code reviews is the secret"  
[X Link](https://x.com/thsottiaux/status/1960946983205789767)  2025-08-28T06:06Z [----] followers, [--] engagements


"We were hesitant to ship it this early but glad to see all the traction and positive feedback already. The adoption has been quite wild. Codex on VSCode copilot (and the UI is so clean 🤌) https://t.co/uFGa6732f7 Codex on VSCode copilot (and the UI is so clean 🤌) https://t.co/uFGa6732f7"  
[X Link](https://x.com/anyuser/status/1961260364840022243)  2025-08-29T02:51Z 29.9K followers, 27.1K engagements


"@gastonmorixe It's a sandboxed vm without internet access you can also get this info by and playing with the container there in the environment setup nothing to hide here https://chatgpt.com/codex https://chatgpt.com/codex"  
[X Link](https://x.com/thsottiaux/status/1963022806201053535)  2025-09-02T23:34Z [----] followers, [--] engagements


"Excited for what we'll build together I'm excited to announce that we're joining OpenAIs Codex team When we started out Xcode had no AI. Building a "Cursor for Xcode" sounded crazy but we managed to do it anyway. And over time we built the best coding agent for iOS & MacOS apps. I'm extremely proud of what we https://t.co/9Wzkixtk9s I'm excited to announce that we're joining OpenAIs Codex team When we started out Xcode had no AI. Building a "Cursor for Xcode" sounded crazy but we managed to do it anyway. And over time we built the best coding agent for iOS & MacOS apps. I'm extremely proud of"  
[X Link](https://x.com/anyuser/status/1963315320980214247)  2025-09-03T18:57Z 29.9K followers, 12.3K engagements


"Exactly why we built deep code review capabilities in codex. 40% of daily code written at Coinbase is AI-generated. I want to get it to 50% by October. Obviously it needs to be reviewed and understood and not all areas of the business can use AI-generated code. But we should be using it responsibly as much as we possibly can. https://t.co/Nmnsdxgosp 40% of daily code written at Coinbase is AI-generated. I want to get it to 50% by October. Obviously it needs to be reviewed and understood and not all areas of the business can use AI-generated code. But we should be using it responsibly as much"  
[X Link](https://x.com/anyuser/status/1963485613539430858)  2025-09-04T06:13Z 29.9K followers, 13.9K engagements


"@yousefrol @TheRealAdamG Yes same underlying coding agent. Some even say it's better than the CLI"  
[X Link](https://x.com/thsottiaux/status/1964742439446745118)  2025-09-07T17:27Z [----] followers, [--] engagements


"@hTrapVader Danke For how to clear codex says: The CLI does not provide a specific clear input shortcut. The fastest way is to use the standard OS shortcut to select all textCtrl + A on Windows/Linux or Cmd + A on macOSand then press Backspace or Delete to remove it"  
[X Link](https://x.com/thsottiaux/status/1964745823776043172)  2025-09-07T17:41Z [----] followers, [----] engagements


"@danielmulec @sama What's wrong with the pro plan"  
[X Link](https://x.com/thsottiaux/status/1964762657279643832)  2025-09-07T18:48Z [----] followers, [----] engagements


"@developerisnow1 What do you use sub-agents for"  
[X Link](https://x.com/thsottiaux/status/1964825637774148025)  2025-09-07T22:58Z [----] followers, [----] engagements


"@cirrus_shakeri For approvals could you give an example of what more granular accept / reject would look like"  
[X Link](https://x.com/thsottiaux/status/1964826117703160061)  2025-09-07T23:00Z [----] followers, [---] engagements


"@tristanbob Are you on windows perhaps More details on sandbox policies and approvals here: https://github.com/openai/codex/blob/main/docs/sandbox.md https://github.com/openai/codex/blob/main/docs/sandbox.md"  
[X Link](https://x.com/thsottiaux/status/1964843869163684271)  2025-09-08T00:10Z [----] followers, [----] engagements


"@goodalexander There is no auto /compact right now indeed. Would you prefer it just kept going and tried to compact the context automatically when it runs out"  
[X Link](https://x.com/thsottiaux/status/1964844401588617445)  2025-09-08T00:13Z [----] followers, [----] engagements


"@bitdeep_ It works with the plus subscription too which is $20 / month"  
[X Link](https://x.com/thsottiaux/status/1964844791138762902)  2025-09-08T00:14Z [----] followers, [----] engagements


"@Hft_freedom Tell me more"  
[X Link](https://x.com/thsottiaux/status/1965277645639496187)  2025-09-09T04:54Z [----] followers, [--] engagements


"@LucasVHoutven Was hotfixed [--] mins ago: https://github.com/openai/codex/releases https://github.com/openai/codex/releases"  
[X Link](https://x.com/thsottiaux/status/1965899225344127214)  2025-09-10T22:04Z [----] followers, [--] engagements


"Sometimes I get stuck here in awe for how pretty @edbayes made this and I forget to talk to codex. You can cycle with cmd+. and I think it's worth installing the extension just for that meditative minute"  
[X Link](https://x.com/anyuser/status/1965914071582326789)  2025-09-10T23:03Z 29.9K followers, 55.2K engagements


"@thebuildguy @edbayes https://marketplace.visualstudio.com/itemsitemName=openai.chatgpt https://marketplace.visualstudio.com/itemsitemName=openai.chatgpt"  
[X Link](https://x.com/thsottiaux/status/1966703208576802888)  2025-09-13T03:19Z [----] followers, [--] engagements


"@JustDoingItBig @embirico We have documentation here: But also are looking to improve this https://github.com/openai/codex/blob/main/docs/authentication.md#connecting-on-a-headless-machine https://github.com/openai/codex/blob/main/docs/authentication.md#connecting-on-a-headless-machine"  
[X Link](https://x.com/thsottiaux/status/1967301005638619428)  2025-09-14T18:54Z [----] followers, [---] engagements


"@enriquemoreno We will keep supporting both gpt-5-codex is further optimized to work great in codex and packs some additional improvements that makes it an even better collaborator. Let me know what you think after trying it"  
[X Link](https://x.com/thsottiaux/status/1967643951663550520)  2025-09-15T17:37Z [----] followers, [----] engagements


"@Conor_D_Dart I recommend using it for two things: a) Helping codex navigate your codebase faster b) Documenting your preference The first one improves efficiency the second one improves how reliable and useful the outcome is to you"  
[X Link](https://x.com/thsottiaux/status/1967707999163801714)  2025-09-15T21:51Z [----] followers, [---] engagements


"After trying gpt-5-codex which one will you keep as your daily driver for coding gpt-5-codex gpt-5 gpt-5-codex gpt-5"  
[X Link](https://x.com/anyuser/status/1968383747105112507)  2025-09-17T18:37Z 29.9K followers, 13.4K engagements


"👀 We are prioritizing improving how the Codex rate limit mechanism works to avoid hitting the issue where you use most of your weekly limit and then hit a [--] day wait period. We also plan to better surface how much is used and when you get close to the limit. Im really confused about the new Codex limits. I was working this morning and suddenly got locked out. Ive never once hit a limit in a 5-hour window and today Ive only sent around [--] messages all on low settings. Same story yesterday only low-setting usage. @thsottiaux https://t.co/rd5gBbD0b2 Im really confused about the new Codex"  
[X Link](https://x.com/anyuser/status/1968689722483466472)  2025-09-18T14:53Z 29.9K followers, 43.3K engagements


"@nifal_adam On pro it is designed for you to not hit limits if you use a single codex instance pretty much continuously"  
[X Link](https://x.com/thsottiaux/status/1968699029035856335)  2025-09-18T15:29Z [----] followers, [----] engagements


"@meekaale You have reached codex . beep . beep . forwarding you to @_catwu"  
[X Link](https://x.com/thsottiaux/status/1968813972741570843)  2025-09-18T23:06Z [----] followers, [---] engagements


"If you could decide this in codex which way would you adjust the dial faster lower limits slower higher limits keep current faster lower limits slower higher limits keep current"  
[X Link](https://x.com/thsottiaux/status/1969061786050580730)  2025-09-19T15:31Z [----] followers, [----] engagements


"@BigForestFarms You can read more about the limits here: https://developers.openai.com/codex/pricing#usage-limits https://developers.openai.com/codex/pricing#usage-limits"  
[X Link](https://x.com/thsottiaux/status/1969257307411202499)  2025-09-20T04:28Z [----] followers, [---] engagements


"@fram3o You can queue up a bunch of "continue" messages but we'll improve here in the future"  
[X Link](https://x.com/thsottiaux/status/1969932693443060168)  2025-09-22T01:12Z [----] followers, [---] engagements


"@ErvistheGreat @LKyyBe @OpenAI @embirico Wdym"  
[X Link](https://x.com/thsottiaux/status/1969932807355928724)  2025-09-22T01:12Z [----] followers, [---] engagements


"This issue is now fixed and gpt-5-codex should be a few percent more context efficient and also never loop with multiple answers showing. Do let us know if you see anything in the behavior that you don't like we're improving it daily. Nothing too scary but found an issue in our stack that leads to slightly higher token usage and sometimes degraded performance for gpt-5-codex where it will cause it to answer multiple times or loop a bit. We'll push out a fix first thing tomorrow and it will feel even better. Nothing too scary but found an issue in our stack that leads to slightly higher token"  
[X Link](https://x.com/anyuser/status/1970253773282165099)  2025-09-22T22:27Z 29.9K followers, 39.9K engagements


"@erd0xbc @AGIGuardian You two should get back to coding"  
[X Link](https://x.com/thsottiaux/status/1970338301627548058)  2025-09-23T04:03Z [----] followers, [---] engagements


"gpt-5-codex is now available in the API + [----] CLI release: visibility into rate limits under /status new /review features auto-compaction and other small improvements"  
[X Link](https://x.com/anyuser/status/1970533973567680936)  2025-09-23T17:01Z 29.9K followers, 195.3K engagements


"@emrahdma /status is the command you are looking for"  
[X Link](https://x.com/thsottiaux/status/1970563118603391275)  2025-09-23T18:57Z [----] followers, [----] engagements


"@velvet_shark This what auto-compaction is for. We also improved /compact to being able to still work even if you reached beyond the context limit with the last message https://github.com/openai/codex/pull/4068 https://github.com/openai/codex/pull/4068"  
[X Link](https://x.com/thsottiaux/status/1970563581575090267)  2025-09-23T18:59Z [----] followers, [----] engagements


"@xqcdp You can set the reasoning effort https://platform.openai.com/docs/guides/reasoning#get-started-with-reasoning https://platform.openai.com/docs/guides/reasoning#get-started-with-reasoning"  
[X Link](https://x.com/thsottiaux/status/1970563945648808039)  2025-09-23T19:00Z [----] followers, [---] engagements


"@iannuttall Release notes here: https://github.com/openai/codex/releases/tag/rust-v0.40.0 https://github.com/openai/codex/releases/tag/rust-v0.40.0"  
[X Link](https://x.com/thsottiaux/status/1970637394581069893)  2025-09-23T23:52Z [----] followers, [----] engagements


"Upgrade to the latest Codex CLI 0.41.0 version to see when limits reset by typing /status. Full release notes at: https://github.com/openai/codex/releases/tag/rust-v0.41.0 @thsottiaux Incredible troll job really. https://t.co/98pjON2mcN https://github.com/openai/codex/releases/tag/rust-v0.41.0 @thsottiaux Incredible troll job really. https://t.co/98pjON2mcN"  
[X Link](https://x.com/anyuser/status/1971298655161417937)  2025-09-25T19:39Z 29.9K followers, 34.1K engagements


"Cute Grok Code just passed [--] trillion tokens Grok Code just passed [--] trillion tokens"  
[X Link](https://x.com/anyuser/status/1973931546550894681)  2025-10-03T02:02Z 29.9K followers, 259K engagements


"@itssidhere Thats called the pro subscription"  
[X Link](https://x.com/thsottiaux/status/1973980515968692281)  2025-10-03T05:16Z [----] followers, [----] engagements


"prompt engineering = tool engineering"  
[X Link](https://x.com/anyuser/status/1974125953497641012)  2025-10-03T14:54Z 29.9K followers, [----] engagements


"New codex CLI [----] dropped tons of little UI improvements to improve readability and feel reduced repetitive approvals for windows support for streamable HTTP MCP servers . $ npm install -g @openai/codex"  
[X Link](https://x.com/anyuser/status/1974174154237882397)  2025-10-03T18:06Z 29.9K followers, 39.8K engagements


"Release notes at https://github.com/openai/codex/releases/tag/rust-v0.44.0 https://github.com/openai/codex/releases/tag/rust-v0.44.0"  
[X Link](https://x.com/thsottiaux/status/1974174206113042485)  2025-10-03T18:06Z [----] followers, [----] engagements


"@John6227647246 @itssidhere You can more than ten times the usage on the pro subscription. For the majority of people with the current state of coding agents it's enough to never be out"  
[X Link](https://x.com/thsottiaux/status/1974228019016343851)  2025-10-03T21:40Z [----] followers, [--] engagements


"@sharrajesh465 Looking for https://marketplace.visualstudio.com/itemsitemName=openai.chatgpt https://marketplace.visualstudio.com/itemsitemName=openai.chatgpt"  
[X Link](https://x.com/thsottiaux/status/1974322926943117702)  2025-10-04T03:57Z [----] followers, [---] engagements


"@Fullmoon211 @embirico There is no routing which model settings are you using"  
[X Link](https://x.com/thsottiaux/status/1976409253251625063)  2025-10-09T22:07Z [----] followers, [--] engagements


"@zakariamehbi @OpenAI @OpenAIDevs @sama Hey Zak what are you seeing Keen to debug and my DMs are open"  
[X Link](https://x.com/thsottiaux/status/1976418926251868317)  2025-10-09T22:46Z [----] followers, [----] engagements


"@iMuffined hmmmmm"  
[X Link](https://x.com/thsottiaux/status/1978684415480852732)  2025-10-16T04:48Z [----] followers, [---] engagements


"@irfanrosandi @sound4movement All the time and believe it or not but we also use ChatGPT"  
[X Link](https://x.com/thsottiaux/status/1979062741236174877)  2025-10-17T05:51Z [----] followers, [--] engagements


"We promised unprecedented transparency for Codex and to take the reports of degradation seriously despite seeing incredible growth week over week. Here is our report and what we have found over the last seven days https://docs.google.com/document/d/1fDJc1e0itJdh0MXMFJtkRiBcxGEFtye6Xc6Ui7eMX4o/editusp=sharing We have concluded our initial investigation into increased reports of degradation for Codex and we will share it out by end of day. What is your bet on what we'll include in there and some of the findings"  
[X Link](https://x.com/anyuser/status/1984465716888944712)  2025-11-01T03:41Z 29.9K followers, 1.1M engagements


"Pro tip: Enable codex code reviews it's been described multiple times to me as superhuman in its ability to spot subtle flaws. Navigate to speedrun the onboarding enable for your repos in the settings. Enjoy deep reviews on every PR. https://chatgpt.com/codex https://chatgpt.com/codex"  
[X Link](https://x.com/anyuser/status/1963012903277412618)  2025-09-02T22:55Z 29.9K followers, 370.7K engagements


"What would you like us to fix in codex What would be the biggest productivity unlock for you and your team"  
[X Link](https://x.com/anyuser/status/1964742011673882691)  2025-09-07T17:26Z 29.9K followers, 242.1K engagements


"@giffmana True noticed this too and we're fixing this"  
[X Link](https://x.com/thsottiaux/status/1964833910392865015)  2025-09-07T23:31Z [----] followers, [----] engagements


"Currently serving gpt-5-codex at degraded speeds about 2X due to the very high demand. We are finding GPUs left and right to match demand we'll keep you posted Demand for GPT-5-Codex is higher than forecasted and we need to stand up more GPUs In the meantime Codex is running 2x slower than targets. Working on fixing this ASAP Demand for GPT-5-Codex is higher than forecasted and we need to stand up more GPUs In the meantime Codex is running 2x slower than targets. Working on fixing this ASAP"  
[X Link](https://x.com/anyuser/status/1967996885500928459)  2025-09-16T16:59Z 29.9K followers, 417.5K engagements


"We have reset everyone's limits for gpt-5-codex to allow you all to benefit more from it today and as a way to make up for the slowdowns earlier while we were provisioning additional GPUs. We'll continue to add capacity throughout the week to maintain the speed. Enjoy"  
[X Link](https://x.com/anyuser/status/1968163721034994139)  2025-09-17T04:02Z 29.9K followers, 748K engagements


"/review now in Codex CLI Use it to get a quick review from codex on your code changes locally. Will expand a lot in coming days send Daniel your feedback Excited to release v1 of /review in the Codex CLI You can now use gpt-5-codex to investigate and find critical bugs in your code. Check it out and let us know if you have any feedback https://t.co/pFuI1XJMg6 Excited to release v1 of /review in the Codex CLI You can now use gpt-5-codex to investigate and find critical bugs in your code. Check it out and let us know if you have any feedback https://t.co/pFuI1XJMg6"  
[X Link](https://x.com/anyuser/status/1968825235328631262)  2025-09-18T23:51Z 29.9K followers, 426.9K engagements


"Team is hard at work on the next version lots of small and big things weve learned from v0. Whats the top thing that you would like to see fixed in GPT-5 Codex OpenAI really cooked with GPT-5 Codex OpenAI really cooked with GPT-5 Codex"  
[X Link](https://x.com/anyuser/status/1973934387323641871)  2025-10-03T02:13Z 29.9K followers, 336.2K engagements


"We've served over 40T tokens for gpt-5-codex since its release less than a month ago. One of our fastest growing models. Cute Cute"  
[X Link](https://x.com/anyuser/status/1975746937925083357)  2025-10-08T02:15Z 29.9K followers, 58.9K engagements


"Codex team is doing a weeklong hackathon starting next week what should we build Other than CLI auto-update we're on that already"  
[X Link](https://x.com/anyuser/status/1976772067791139082)  2025-10-10T22:09Z 29.9K followers, 205.4K engagements


"We are actively working on making codex faster and we will likely ship most of these as incremental improvements. For example Codex CLI v0.45 included a change that made it 10% more token efficient at equal result. I'm bottlenecked by Codex. Every day I have to choose [--] or [--] things to work on because it will spend a long time on each. My routine is like: - wake up - write a prompt - send it to Codex - go eat something - come back it is still not done - take a shower - come back still I'm bottlenecked by Codex. Every day I have to choose [--] or [--] things to work on because it will spend a long"  
[X Link](https://x.com/anyuser/status/1978250242639134991)  2025-10-15T00:03Z 29.9K followers, 80.8K engagements


"Codex is now natively supported in Zed. What's extra exciting is that all of it is open source Zed itself ACP and the Codex Agent + SDK which live under the openai/codex oss repo. https://zed.dev/blog/codex-is-live-in-zed https://zed.dev/blog/codex-is-live-in-zed"  
[X Link](https://x.com/anyuser/status/1978969611346546958)  2025-10-16T23:41Z 29.9K followers, 29.7K engagements


"Tuning the GPUs. For codex on the plus subscription ($20 / month) would you prefer to be able to perform more overall work but a bit slower . or less overall work but it gets done faster higher limits bit slower lower limits bit faster show me the results no change higher limits bit slower lower limits bit faster show me the results no change"  
[X Link](https://x.com/anyuser/status/1979590503764307998)  2025-10-18T16:48Z 29.9K followers, 10.8K engagements


"Bit hidden but powerful feature in Codex CLI: custom prompts aka custom slash commands. Just improved the documentation for it and added an example of one I use all the time /prompts:draftpr https://github.com/openai/codex/blob/main/docs/prompts.md https://github.com/openai/codex/blob/main/docs/prompts.md"  
[X Link](https://x.com/anyuser/status/1979692123340542357)  2025-10-18T23:32Z 29.9K followers, 20.4K engagements


"Codex CLI supports global instructions under /.codex/AGENTS.md which is useful for things you always want Codex to take into account. We've updated the documentation to more clearly explain this and the precedence rules at https://github.com/openai/codex/blob/main/docs/agents_md.md http://AGENTS.md https://github.com/openai/codex/blob/main/docs/agents_md.md http://AGENTS.md"  
[X Link](https://x.com/anyuser/status/1980178813817618512)  2025-10-20T07:46Z 29.9K followers, 86.7K engagements


"In Codex CLI we've had occasional occurrences of "unsupported model" mid session which was annoying. Team has found the bug and we're rolling out the fix. We'll continue to improve reliability and performance over time"  
[X Link](https://x.com/anyuser/status/1980434010770948529)  2025-10-21T00:40Z 29.9K followers, 35.3K engagements


"This is now fixed together with an improvement that means we don't interrupt mid-turn even if you hit your limits. May the Codex sessions be smooth and full of wonder. In Codex CLI we've had occasional occurrences of "unsupported model" mid session which was annoying. Team has found the bug and we're rolling out the fix. We'll continue to improve reliability and performance over time. https://t.co/w39MYViVwT In Codex CLI we've had occasional occurrences of "unsupported model" mid session which was annoying. Team has found the bug and we're rolling out the fix. We'll continue to improve"  
[X Link](https://x.com/anyuser/status/1980697275543544033)  2025-10-21T18:06Z 29.9K followers, 17.9K engagements


"Were working on better windows support for Codex please chime in Codex IDE Extension users on Windows: We have a few improvements coming that we'd love help testing. Today's prerelease v0.5.22 supports using WSL to run commands in bash for better performance and fewer approvals with sandboxing. Test instructions in 🧵 https://t.co/1VmsdOmYiG Codex IDE Extension users on Windows: We have a few improvements coming that we'd love help testing. Today's prerelease v0.5.22 supports using WSL to run commands in bash for better performance and fewer approvals with sandboxing. Test instructions in 🧵"  
[X Link](https://x.com/anyuser/status/1980855994528985533)  2025-10-22T04:37Z 29.9K followers, [----] engagements


"Codex hackathon over. Counted not less than [--] demos presented today team will be locked in over coming weeks to bring you the best of it. Never been more excited for the future of software Codex team is doing a weeklong hackathon starting next week what should we build Other than CLI auto-update we're on that already Codex team is doing a weeklong hackathon starting next week what should we build Other than CLI auto-update we're on that already"  
[X Link](https://x.com/anyuser/status/1980864822737854527)  2025-10-22T05:12Z 29.9K followers, 28.1K engagements


"The cycle of growth and additional GPUs continues. Proud of the team continuously dealing with new peaks of traffic and pushing to improve codex every day"  
[X Link](https://x.com/anyuser/status/1981594505464975635)  2025-10-24T05:32Z 29.9K followers, [----] engagements


"Codex CLI [----] is here. Packs many quality of life improvements MCP features stabler /compact and auto-compaction better documentation --add-dir for additional working directories. Lots of contributions from the community thank you. Full notes: https://github.com/openai/codex/releases/tag/rust-v0.48.0 https://github.com/openai/codex/releases/tag/rust-v0.48.0"  
[X Link](https://x.com/anyuser/status/1981775470367846898)  2025-10-24T17:31Z 29.9K followers, 159.9K engagements


"npm install -g @openai/codex"  
[X Link](https://x.com/thsottiaux/status/1981775536469967245)  2025-10-24T17:31Z 14.8K followers, [----] engagements


"@Conor_D_Dart Not yet team is hard at work improving windows support"  
[X Link](https://x.com/thsottiaux/status/1981789368064127005)  2025-10-24T18:26Z [----] followers, [----] engagements


"New Codex CLI version [----] with /feedback command. Use it liberally to report bugs model issues or send us feature requests in the moment. Let the flywheel begin"  
[X Link](https://x.com/anyuser/status/1982339555723550801)  2025-10-26T06:52Z 29.9K followers, 31.4K engagements


"Which features do you consider bloat in codex @thsottiaux Feedback commands only work if users actually encounter problems worth reporting. Most CLI tools die from feature bloat not lack of input channels. @thsottiaux Feedback commands only work if users actually encounter problems worth reporting. Most CLI tools die from feature bloat not lack of input channels"  
[X Link](https://x.com/thsottiaux/status/1982468838265651628)  2025-10-26T15:26Z 11.6K followers, 13.1K engagements


"@vinay_sirohiya Please feel free to use the /feedback command is particularly useful to report occasional model mishaps and includes diagnostics data that is helpful to us https://github.com/openai/codex/issues https://github.com/openai/codex/issues"  
[X Link](https://x.com/thsottiaux/status/1982703071521292764)  2025-10-27T06:57Z [----] followers, [---] engagements


"On While puzzling we are taking this seriously and are in the middle of a full investigation started last Friday. Current approach: 1) Upgrades to /feedback command 2) Reduce surface of things that could cause issues 3) Evals and qualitative checks https://x.com/i/trending/1983562047729045942 https://x.com/i/trending/1983562047729045942"  
[X Link](https://x.com/anyuser/status/1983675985641009260)  2025-10-29T23:23Z 29.9K followers, 278.9K engagements


"@AtlantisPleb You might be onto something here"  
[X Link](https://x.com/thsottiaux/status/1983724780919648360)  2025-10-30T02:36Z 11.3K followers, [----] engagements


"Lots of you asked for a $100 plan for Codex but we think our new credits approach might be better and you can now get more out of Codex based on what you actually need. If you want more Codex than Plus but dont have $200 to subscribe to Pro we have a new option for you: Credits They also work if you're maxing out your Pro limits. Keep the feedback coming as always. PS In celebration we reset Codex rate limits for everyone 🎉 If you want more Codex than Plus but dont have $200 to subscribe to Pro we have a new option for you: Credits They also work if you're maxing out your Pro limits. Keep"  
[X Link](https://x.com/anyuser/status/1983973493864894739)  2025-10-30T19:05Z 29.9K followers, 35.1K engagements


"Codex CLI v [----] is out. Lots of polish further reduction in crashes that would happen 0.1% of the time execute commands directly with cmd and new /undo command"  
[X Link](https://x.com/anyuser/status/1983986186751467857)  2025-10-30T19:55Z 29.9K followers, 37.5K engagements


"Full notes: https://github.com/openai/codex/releases/tag/rust-v0.52.0 https://github.com/openai/codex/releases/tag/rust-v0.52.0"  
[X Link](https://x.com/thsottiaux/status/1983986230531703121)  2025-10-30T19:55Z 11.9K followers, [----] engagements


"We have concluded our initial investigation into increased reports of degradation for Codex and we will share it out by end of day. What is your bet on what we'll include in there and some of the findings"  
[X Link](https://x.com/anyuser/status/1984351191527276631)  2025-10-31T20:06Z 29.9K followers, 300.1K engagements


"If you use Codex on Windows this is for you. We're investing in native sanboxing and have published our first experimental version as part of v0.53 please give us feedback as we continue to build this Codex Experimental Windows Sandbox 📣 CLI v0.53 has a new highly experimental filesystem and network sandbox that we'd love help testing. Once it's stable Codex agent mode will run more securely and safely with far fewer approvals. The current biggest caveat is that it does Codex Experimental Windows Sandbox 📣 CLI v0.53 has a new highly experimental filesystem and network sandbox that we'd love"  
[X Link](https://x.com/anyuser/status/1984670412492718455)  2025-11-01T17:14Z 29.9K followers, 33.1K engagements


"Bit of a breakthrough today in how we can do better to provide higher and more predictable usage for Codex within your ChatGPT subscription. Expect some nice updates here in the coming two weeks"  
[X Link](https://x.com/anyuser/status/1985605628770058501)  2025-11-04T07:10Z 29.9K followers, 33.4K engagements


"@CallofdutyFan32 Thanks for using /feedback that's perfect. Would love a community showcase and I find twitter vibes quite nice lately. The contrast with reddit is interesting"  
[X Link](https://x.com/thsottiaux/status/1985616411608432935)  2025-11-04T07:53Z 13.9K followers, [---] engagements


"We are seeing elevated system errors for codex and are actively working to resolve"  
[X Link](https://x.com/anyuser/status/1986157620303176048)  2025-11-05T19:44Z 29.9K followers, 11.8K engagements


"Codex outage resolved. We suffered a minor outage for last [--] minutes which turned into a global outage for last [--] minutes this is now resolved and we have reset the rate limits. Please let me know if you still see issues"  
[X Link](https://x.com/anyuser/status/1986166501435711936)  2025-11-05T20:19Z 29.9K followers, 63.5K engagements


"Edit: This is resolved. Original post: We are seeing elevated system errors for codex and are actively working to resolve"  
[X Link](https://x.com/anyuser/status/1986166900322361810)  2025-11-05T20:21Z 29.9K followers, [----] engagements


"@nummanthinks @OpenAI Wrong tibo"  
[X Link](https://x.com/anyuser/status/1986581833790996888)  2025-11-06T23:49Z 29.9K followers, [----] engagements


"Small upgrade to Codex we have updated gpt-5-codex within Codex. The model should feel better more collaborative and across the board we gained a few percentage points for the evals we care about. Also a tad more token-efficient needing 3% less to achieve similar results. Is it me or Codex got a little smarter today Same gpt-5-codex high on v0.55 but now it's more careful about edge cases and I don't have have to guide it too much. GG @thsottiaux @embirico Is it me or Codex got a little smarter today Same gpt-5-codex high on v0.55 but now it's more careful about edge cases and I don't have"  
[X Link](https://x.com/anyuser/status/1986602121572327650)  2025-11-07T01:10Z 29.9K followers, 254.3K engagements


"More predictable usage in Codex. We landed optimizations behind the scenes that mean that you now will get the same usage throughout the day irrespective of the load or how traffic is routed. Before you could get unlucky and hit a few cache misses in a row and get less usage"  
[X Link](https://x.com/anyuser/status/1986695544887255104)  2025-11-07T07:21Z 29.9K followers, 14.8K engagements


"We gave gpt-5-codex a small sibling you can now try gpt-5-codex-mini in Codex. Smaller so up to 4X more usage included. $ codex -m gpt-5-codex-mini Also increased usage limits by 50% for Plus Edu and Team plans and gave Pro priority processing for a faster experience. You can now get more Codex usage from your plan and credits with three updates today: [--] GPT-5-Codex-Mini a more compact and cost-efficient version of GPT-5-Codex [--] 50% higher rate limits for ChatGPT Plus Business and Edu [--] Priority processing for ChatGPT Pro and You can now get more Codex usage from your plan and credits with"  
[X Link](https://x.com/anyuser/status/1986863197803192782)  2025-11-07T18:27Z 29.9K followers, 90.8K engagements


"@GlitchInTheCode Upgrade the CLI it's in [----] https://github.com/openai/codex/releases/tag/rust-v0.56.0 https://github.com/openai/codex/releases/tag/rust-v0.56.0"  
[X Link](https://x.com/thsottiaux/status/1986868265742966985)  2025-11-07T18:48Z 12.4K followers, [----] engagements


"@Quirk2Muffin You don't have to just use the codex CLI or vscode extension $ npm install -g @openai/codex it works entirely on your local machine and filesystem"  
[X Link](https://x.com/thsottiaux/status/1986879072002261189)  2025-11-07T19:30Z 12.4K followers, [----] engagements


"Image inputs are currently not working with GPT-5-Codex Mini we are working on fixing this"  
[X Link](https://x.com/anyuser/status/1987222166417645601)  2025-11-08T18:14Z 29.9K followers, 51.6K engagements


"@embirico @GregKonush @OpenAI https://x.com/thsottiaux/status/1987222166417645601 Image inputs are currently not working with GPT-5-Codex Mini we are working on fixing this https://x.com/thsottiaux/status/1987222166417645601 Image inputs are currently not working with GPT-5-Codex Mini we are working on fixing this"  
[X Link](https://x.com/thsottiaux/status/1987222727787508126)  2025-11-08T18:16Z 12.3K followers, [---] engagements


"Image inputs are now supported with GPT-5-Codex Mini. You can try it with $ codex -m gpt-5-codex-mini Image inputs are currently not working with GPT-5-Codex Mini we are working on fixing this Image inputs are currently not working with GPT-5-Codex Mini we are working on fixing this"  
[X Link](https://x.com/anyuser/status/1987255184331055514)  2025-11-08T20:25Z 29.9K followers, 31.1K engagements


"@davehelmuth Natively multimodal. Same as GPT-5 and GPT-5-Codex"  
[X Link](https://x.com/thsottiaux/status/1987267762394402872)  2025-11-08T21:15Z 12.7K followers, [---] engagements


"We've made changes that allows us to serve GPT-5-Codex-Mini at twice the number of tokens per second compared to last Friday which means it should feel about twice as fast for tasks rely heavily on generating a lot of code. $ codex -m gpt-5-codex-mini"  
[X Link](https://x.com/anyuser/status/1988100565071786125)  2025-11-11T04:24Z 29.9K followers, 102K engagements


"We've upgraded all our models in Codex. You can now use gpt-5.1 gpt-5.1-codex and gpt-5.1-codex-mini. All are a step-up that you should able to feel in daily use and you can expect us to continue pushing here on the frontier sooner than you might think"  
[X Link](https://x.com/anyuser/status/1989060822153465873)  2025-11-13T20:00Z 29.9K followers, 74.7K engagements


"PSA: We changed usage display across Codex CLI VS Code extension and web to be consistent and it now displays usage *remaining*. Like fuel in a car. If it's close to 100% it's good if it's close to 0% go refuel. @thsottiaux So far it's not disappointing. But please. never again invert a progress bar silently. 😂 https://t.co/2ZcnMZuzYP @thsottiaux So far it's not disappointing. But please. never again invert a progress bar silently. 😂 https://t.co/2ZcnMZuzYP"  
[X Link](https://x.com/anyuser/status/1989080994775986329)  2025-11-13T21:20Z 29.9K followers, 46.2K engagements


"@normal_evan The version we shipped last week was an earlier version of gpt-5.1-codex and now with the full launch I am quite confident that you will like it just the same if not a bit more https://x.com/thsottiaux/status/1986602121572327650 Small upgrade to Codex we have updated gpt-5-codex within Codex. The model should feel better more collaborative and across the board we gained a few percentage points for the evals we care about. Also a tad more token-efficient needing 3% less to achieve similar results. https://x.com/thsottiaux/status/1986602121572327650 Small upgrade to Codex we have"  
[X Link](https://x.com/thsottiaux/status/1989093492967240103)  2025-11-13T22:10Z 14.5K followers, [----] engagements


"Continuous incremental improvements are often underrated. But looking back it's surprising how much of a jump we've gone through in less than two months since the release of gpt-5-codex. GPT-5.1 Codex beats Sonnet [---] Thinking on SWE-Bench while being [--] times cheaper ouch https://t.co/5RZdaMIzCZ GPT-5.1 Codex beats Sonnet [---] Thinking on SWE-Bench while being [--] times cheaper ouch https://t.co/5RZdaMIzCZ"  
[X Link](https://x.com/anyuser/status/1989533615236264029)  2025-11-15T03:19Z 29.9K followers, 40.6K engagements


"Imagine how unhobbled Codex will be once we rollout a proper fix to this. In the meantime believe me there are . technical reasons for why this is. Not that I like them. This is even funnier. Codex will truncate any Bash/MCP tool output to [---] lines or 10kb. If the tool call outputs more than that then the model only gets to see the first [---] lines and the last [---] lines but nothing in the middle. Been like that since August. No wonder poor https://t.co/1kdkiraRjk This is even funnier. Codex will truncate any Bash/MCP tool output to [---] lines or 10kb. If the tool call outputs more than that"  
[X Link](https://x.com/anyuser/status/1989940347494084683)  2025-11-16T06:15Z 29.9K followers, 351.1K engagements


"@sdmat123 Reason: Precise tokenization adds latency specific tokenizer we use in production is not (yet) opensource and unbounded reads aren't good either as they can blow up the context so you have to limit in some way"  
[X Link](https://x.com/anyuser/status/1989947362136260787)  2025-11-16T06:43Z 29.9K followers, 86K engagements


"@AtlantisPleb Seriously thinking about it but rewriting from scratch is rarely the solution to your problem"  
[X Link](https://x.com/thsottiaux/status/1989948198253990046)  2025-11-16T06:46Z 13.6K followers, [----] engagements


"We are really bad at naming but I'm excited for the rest of this week. Codex team is on fire"  
[X Link](https://x.com/anyuser/status/1990945361687687497)  2025-11-19T00:48Z 29.9K followers, 243.9K engagements


"I think max is a good name for this little one"  
[X Link](https://x.com/anyuser/status/1991038838966530156)  2025-11-19T07:00Z 29.9K followers, 33.9K engagements


"New model is out in Codex. Gets to same quality of solution faster and raises the ceiling for how complex of a tasks are achievable. $ codex -m gpt-5.1-codex-max Best experienced in the latest CLI version [----] which also packs a lot of other fixes and improvements"  
[X Link](https://x.com/anyuser/status/1991210545253609875)  2025-11-19T18:22Z 29.9K followers, 200.1K engagements


"https://openai.com/index/gpt-5-1-codex-max/ https://openai.com/index/gpt-5-1-codex-max/"  
[X Link](https://x.com/anyuser/status/1991210571489034394)  2025-11-19T18:22Z 29.9K followers, [----] engagements


"Codex CLI with --search now works with gpt-5.1-codex-max again. We've also landed the model in the Codex VS Code extension which is the daily driver for a lot of users and have more updates coming that will further improve the experience this week. We are seeing a small percentage of Codex sessions fail with an error related to input..id team is working on fixing. We are seeing a small percentage of Codex sessions fail with an error related to input..id team is working on fixing"  
[X Link](https://x.com/anyuser/status/1991289561339818106)  2025-11-19T23:36Z 29.9K followers, 21.6K engagements


"We have rolled out a fix to a model issue which could cause gpt-5.1-codex-max to loop. Thank you for using /feedback command to report. Also shipped additional improvements in v0.61 of the Codex CLI and which fixes an issue where the model would sometimes struggle with quoting"  
[X Link](https://x.com/thsottiaux/status/1991733487355982089)  2025-11-21T05:00Z 13.9K followers, 18.4K engagements


"for the release notes and fix for quoting issues in some shells is https://github.com/openai/codex/pull/6948 https://github.com/openai/codex/releases/tag/rust-v0.61.0 https://github.com/openai/codex/pull/6948 https://github.com/openai/codex/releases/tag/rust-v0.61.0"  
[X Link](https://x.com/thsottiaux/status/1991733622559371580)  2025-11-21T05:01Z 14.2K followers, [----] engagements


"@kcosr Should be fixed in v0.61 please upgrade https://github.com/openai/codex/releases/tag/rust-v0.61.0 https://github.com/openai/codex/releases/tag/rust-v0.61.0"  
[X Link](https://x.com/thsottiaux/status/1991735192143487475)  2025-11-21T05:07Z 13.9K followers, [----] engagements


"The gpt-5.1-codex-max model is significantly more of a step change than the name might imply and it will take us some days/weeks to allow to fully express itself in Codex in a way that is out of the box easy for everyone"  
[X Link](https://x.com/thsottiaux/status/1991761542690533611)  2025-11-21T06:52Z 14.6K followers, 50.6K engagements


"We have deployed in-place changes that make GPT-5.1-Codex-Max a more reliable coding partner and reduces the occurrence of one in one thousand type of odd behavior such as running a spurious echo or checking the git status"  
[X Link](https://x.com/thsottiaux/status/1992021391416459601)  2025-11-22T00:04Z 21.4K followers, 26.5K engagements


"Codex is suffering from a partial outage since 10:30am PST team is working to bring back service"  
[X Link](https://x.com/thsottiaux/status/1992312110181990407)  2025-11-22T19:19Z 14.3K followers, 46.3K engagements


"Only affecting traffic in USA with traffic in rest of the world unaffected (unfortunately they are either sleeping or it's late and it's in the middle of the day for the USA)"  
[X Link](https://x.com/thsottiaux/status/1992320096065442008)  2025-11-22T19:51Z 14.2K followers, [----] engagements


"Codex is back and operating at normal latency for users globally. The outage affected a significant part of our users who were in the US or routed to US clusters for processing. Team is making immediate improvements to help ensure Codex can run 24/7 in the future. Codex is suffering from a partial outage since 10:30am PST team is working to bring back service. Codex is suffering from a partial outage since 10:30am PST team is working to bring back service"  
[X Link](https://x.com/thsottiaux/status/1992365535636656624)  2025-11-22T22:52Z 14.5K followers, 65.2K engagements


"We have reset rate limits for all Codex users to compensate for the unusual high latencies in or near the US in the previous hours. Grateful for the team working hard behind the scenes to keep Codex running 24/7. I do recommend trying out GPT-5.1-Codex-Max too if you havent Codex is back and operating at normal latency for users globally. The outage affected a significant part of our users who were in the US or routed to US clusters for processing. Team is making immediate improvements to help ensure Codex can run 24/7 in the future. Codex is back and operating at normal latency for users"  
[X Link](https://x.com/thsottiaux/status/1992370994028388670)  2025-11-22T23:13Z 14.5K followers, 47.6K engagements


"@Aristotleless These were given out for free to users of Codex web and had an expiry date that ended last week. I think we communicated this pretty clearly"  
[X Link](https://x.com/thsottiaux/status/1992373982339994095)  2025-11-22T23:25Z 13.9K followers, [---] engagements


"@dusan_cv A lot of our investments in making codex great natively on Windows came together last week. Glad its working better for you"  
[X Link](https://x.com/thsottiaux/status/1992374801273610623)  2025-11-22T23:29Z 14.3K followers, [----] engagements


"@manabiSRS I understand what you mean. You would rather it refilled until the reset time and then you get a reset again instead of getting a new week of usage. Will look into this but no promise"  
[X Link](https://x.com/thsottiaux/status/1992381761788211407)  2025-11-22T23:56Z 14.3K followers, [----] engagements


"Seeing GPT-5.1-Codex-Max be first in terminal bench [---] is nice because it is within the Codex CLI harness which means it measures something that is readily usable by all with a $20 ChatGPT subscription without tricks. It also makes me very excited for the weeks to come"  
[X Link](https://x.com/thsottiaux/status/1993203428324159683)  2025-11-25T06:21Z 16.4K followers, 29.3K engagements


"Always improving welcome back Opus [---] too expensive Gemini [--] rate limits are just non-existent (and it's lazy to the point that it's a pain to work with). Welcome back. Codex Opus [---] too expensive Gemini [--] rate limits are just non-existent (and it's lazy to the point that it's a pain to work with). Welcome back. Codex"  
[X Link](https://x.com/thsottiaux/status/1993507861574697030)  2025-11-26T02:31Z 14.5K followers, 46.2K engagements


"Codex team is working on a few experimental projects that are starting to shape up and Im excited to share more about soon. But Im curious what would you like to see ship or improved by the end of the year other than better models"  
[X Link](https://x.com/thsottiaux/status/1993508631015571723)  2025-11-26T02:34Z 18K followers, 212.8K engagements


"@iannuttall @dkundel @steipete For faster way to search models what do you think about saved presets"  
[X Link](https://x.com/thsottiaux/status/1993569577188692195)  2025-11-26T06:36Z 14.3K followers, [----] engagements


"Asked Codex for help on this one. After it scraped all the replies there were over [---] unique suggestions which it then categorized + prioritized and I re-shared with the team. Lots of these we're hoping to ship or improve before end of year. Thanks everyone Codex team is working on a few experimental projects that are starting to shape up and Im excited to share more about soon. But Im curious what would you like to see ship or improved by the end of the year other than better models Codex team is working on a few experimental projects that are starting to shape up and Im excited to share"  
[X Link](https://x.com/thsottiaux/status/1995774702187692391)  2025-12-02T08:39Z 14.6K followers, 41.5K engagements


"We have discovered an issue in Codex that could lead to seeing inconsistent usage limits being displayed across Web CLI and VS Code IDE. Team is working on fixing"  
[X Link](https://x.com/thsottiaux/status/1995982616538087875)  2025-12-02T22:25Z 16K followers, 42.2K engagements


"We have reset the usage limits for everyone in Codex which helps us mitigate the issue the fastest and the team will continue to work on the underlying cause. This should propagate to all users over the next [--] mins enjoy and apologies for the confusion We have discovered an issue in Codex that could lead to seeing inconsistent usage limits being displayed across Web CLI and VS Code IDE. Team is working on fixing. We have discovered an issue in Codex that could lead to seeing inconsistent usage limits being displayed across Web CLI and VS Code IDE. Team is working on fixing"  
[X Link](https://x.com/thsottiaux/status/1995988609896513743)  2025-12-02T22:49Z 16K followers, 47K engagements


"@akmatori @sama @OpenAI yw still in experimental phase right now"  
[X Link](https://x.com/thsottiaux/status/1995988758886580349)  2025-12-02T22:49Z 14.5K followers, [----] engagements


"@youyuxi @OpenAI As long as they don't add a if (isOpenAI()) slow(); we should be fine I think"  
[X Link](https://x.com/thsottiaux/status/1996108219744485669)  2025-12-03T06:44Z 14.6K followers, 30.2K engagements


"@bheemreddy181 @youyuxi @OpenAI How would this work beyond the current message queueing feature where you can type and submit a prompt and it gets submitted after the turn completes"  
[X Link](https://x.com/thsottiaux/status/1996112415843647840)  2025-12-03T07:01Z 14.5K followers, [----] engagements


"GPT-5.1-Codex-Max is now in the API and we shipped Codex CLI v.065 with lots of quality of life improvements (better resume tooltips screen display performance and stability) $ npm i -g @openai/codex We just shipped GPT-5.1-Codex-Max to API Codex agents everywhere are now faster & smarter whether you're using Codex in the CLI via API key or in other tools like @cursor_ai and @code. More about how Cursor updated their harness to make the most of it https://t.co/FSHuC55kU1 We just shipped GPT-5.1-Codex-Max to API Codex agents everywhere are now faster & smarter whether you're using Codex in the"  
[X Link](https://x.com/thsottiaux/status/1996664512142073990)  2025-12-04T19:34Z 16K followers, 46.3K engagements


"What should we name the next Codex model I'm not sure how we can top GPT-5.1-Codex-Max with Extra High reasoning. Does the name matter or is all about capabilities"  
[X Link](https://x.com/thsottiaux/status/1997144906570387535)  2025-12-06T03:23Z 16K followers, 103.8K engagements


"Incredible to have GPT-5.2 out. Available in Codex and best enjoyed after upgrading to v0.71. Back to short names. at least for a bit $ codex -m gpt-5.2 GPT-5.2 Thinking evals https://t.co/Kcnz3ZIwye GPT-5.2 Thinking evals https://t.co/Kcnz3ZIwye"  
[X Link](https://x.com/thsottiaux/status/1999208867390275707)  2025-12-11T20:05Z 16.2K followers, 41.4K engagements


"GPT-5.2 is more popular that we anticipated in Codex and we are adding more compute to keep up and keep the latency where we want. The night will be fun"  
[X Link](https://x.com/thsottiaux/status/1999346752751829190)  2025-12-12T05:13Z 16.4K followers, 58.4K engagements


"Having a creative and inquisitive mindset is quickly becoming the bottleneck. And I see top codex users of codex ask more questions rather than coming with a preconceived idea of what or how something should be done. I've been testing GPT-5.2 for a while now. It's a serious leap forward in complex reasoning math coding and simulations. It built a full 3D graphics engine in a single file. Interactive controls. 4K export. One shot. The pace of progress is unreal. https://t.co/yTaEA9Dbyy I've been testing GPT-5.2 for a while now. It's a serious leap forward in complex reasoning math coding and"  
[X Link](https://x.com/thsottiaux/status/1999744205829341366)  2025-12-13T07:32Z 16.2K followers, 17.9K engagements


"@LucasVHoutven Lots of us just plan by collaborating on a It works well and my general rule for these things is to strive for generality and composability. However we might still do something here to have it be good out of the box. http://plan.md http://plan.md"  
[X Link](https://x.com/thsottiaux/status/1999747742558748697)  2025-12-13T07:46Z 15.4K followers, [---] engagements


"Ability to use long context effectively directly translates in better and more reliable results for long tasks on Codex. Users like to /compact early or keep their context used small but with GPT-5.2 this isnt needed as much. https://t.co/pND1qGFgDk While we are all excited by the concept of larger context windows it's important to evaluate the QUALITY of those context windows. "In practical terms this enables professionals to use GPT5.2 to work with long documentssuch as reports contracts https://t.co/m97pDSu69t https://t.co/pND1qGFgDk While we are all excited by the concept of larger"  
[X Link](https://x.com/thsottiaux/status/1999874362162491732)  2025-12-13T16:09Z 16.3K followers, 31.6K engagements


"@xMysteryLC We keep the rest for output tokens. So 256k (max input budget) + 144k (min output budget)"  
[X Link](https://x.com/thsottiaux/status/1999879198933213287)  2025-12-13T16:28Z 16.2K followers, [---] engagements


"@the404coder What should we improve"  
[X Link](https://x.com/thsottiaux/status/1999879384476631061)  2025-12-13T16:29Z 16.5K followers, [---] engagements


"Codex team is hiring. We are rapidly improving the Codex agent but want to build more delight and have a highly ambitious roadmap. Share your cool fork of our open source codex repo or similarly impressive rust oss project. Opportunities for full time and contracting"  
[X Link](https://x.com/thsottiaux/status/1999980696534909144)  2025-12-13T23:12Z 16.8K followers, 237.7K engagements


"@iMuffined Yes"  
[X Link](https://x.com/thsottiaux/status/1999987773311492429)  2025-12-13T23:40Z 16.1K followers, [----] engagements


"We have added *experimental* support for skills and it combines well with GPT-5.2. Already seeing some cool things in the wild that leverage skills in codex. I think about skills as an extension of with progressive disclosure. $ codex --enable skills http://AGENTS.md codex + skills finetunes Qwen3-0.6B to +6 on humaneval and beats the base score on the first run. I reran the experiment from this week but used codex's new skills integration. Like claude code codex consumes the full skill into context and doesn't start with failing runs. https://t.co/QyClw9qWyO http://AGENTS.md codex + skills"  
[X Link](https://x.com/thsottiaux/status/2000295405242966334)  2025-12-14T20:02Z 16.5K followers, 116K engagements


"Max was a bit lonely. Fun to play with GPT-Image [---] to give Max a cute caribou friend"  
[X Link](https://x.com/thsottiaux/status/2001472364329251070)  2025-12-18T01:59Z 17.7K followers, 23.3K engagements


"@randomdude22401 What are your thoughts on 5.2"  
[X Link](https://x.com/thsottiaux/status/2001494367094346104)  2025-12-18T03:26Z 16.3K followers, [---] engagements


"@yacineMTB Love this feedback thank you. We intend to make continued quick progress here and will be curious if today moves things in the right direction for you"  
[X Link](https://x.com/thsottiaux/status/2001670095387701373)  2025-12-18T15:05Z 16.3K followers, [----] engagements


"GPT-5.2-Codex is out further advancing our SoTA for professional software engineering and long-running agentic coding work. It improves on instruction following long-context understanding and pushes the frontier including on cyber. $ codex -m gpt-5.2-codex https://openai.com/index/introducing-gpt-5-2-codex/ https://openai.com/index/introducing-gpt-5-2-codex/"  
[X Link](https://x.com/thsottiaux/status/2001720483872674269)  2025-12-18T18:25Z 23K followers, 95.5K engagements


"@Yampeleg Let's start with GPT-5.2-Codex https://openai.com/index/introducing-gpt-5-2-codex/ https://openai.com/index/introducing-gpt-5-2-codex/"  
[X Link](https://x.com/thsottiaux/status/2001722403605635074)  2025-12-18T18:33Z 17.8K followers, 26.4K engagements


"I am biased obviously but I like the terseness of our models and GPT-5.2-Codex also won't hesitate to correct you if you are clearly wrong in your assumptions. But I am curious what do you all prefer here and what do you want to see improved personality / collaboration wise"  
[X Link](https://x.com/thsottiaux/status/2001730153744121931)  2025-12-18T19:03Z 21.4K followers, 26.9K engagements


"@Gusarich /feedback and we'll take a look"  
[X Link](https://x.com/thsottiaux/status/2001758350976127463)  2025-12-18T20:55Z 16.4K followers, [---] engagements


"Codex doesnt have a skill issue anymore. The feature is out of experimental and now generally available in CLI and IDE extension. Thanks for all the feedback when it was experimental Codex now officially supports skills Skills are reusable bundles of instructions scripts and resources that help Codex complete specific tasks. You can call a skill directly with $.skill-name or let Codex choose the right one based on your prompt. https://t.co/evbKtMkhaU Codex now officially supports skills Skills are reusable bundles of instructions scripts and resources that help Codex complete specific tasks."  
[X Link](https://x.com/thsottiaux/status/2002138008913584215)  2025-12-19T22:04Z 17.8K followers, 55.3K engagements


"@AAAzzam Skills are an extension of with the only difference lying in the progressive disclosure of the content and the added structure that makes it easier for us humans to manage the content. http://AGENTS.md http://AGENTS.md"  
[X Link](https://x.com/thsottiaux/status/2002491958254825549)  2025-12-20T21:31Z 17.8K followers, 44K engagements


"GPT-5.2-Codex is now by and far the most popular model in Codex. Production distribution shows significant token-efficiency gain at equal reasoning effort compared to GPT-5.2 in Codex. Will share plots soon. What's your experience"  
[X Link](https://x.com/thsottiaux/status/2002495277996061037)  2025-12-20T21:44Z 17.8K followers, 87.2K engagements


"One last little present from the Codex team. Today we are launching our latest model: GPT-5.2-Codex-XMas. It performs exactly the same as GPT-5.2-Codex just with a needed personality upgrade for the season. Enjoy coding with Santa Codex $ codex -m gpt-5.2-codex-xmas For Codex users to thank you all for the fun we've had over the last months our first gift is that we have **reset rate limits and are lifting the usage limits to 2X the usual limits until the 1st of Jan**. Wish you all a merry holiday and lots of coding Spread the word 👀 https://t.co/uKAZLhpp7r For Codex users to thank you all"  
[X Link](https://x.com/thsottiaux/status/2004315133225701847)  2025-12-25T22:15Z 21.1K followers, 377.8K engagements


"Limited time only after December it will have to go rest"  
[X Link](https://x.com/thsottiaux/status/2004315631836189090)  2025-12-25T22:17Z 17.8K followers, 10.8K engagements


"Models have improved so much over the last recent months that Im convinced that we need a new way of interfacing to fully benefit from the capability jump. I expect we will see very fast evolution of the model human interaction layer over the coming months"  
[X Link](https://x.com/thsottiaux/status/2004784062830186687)  2025-12-27T05:19Z 18.7K followers, 42.1K engagements


"When the feature is enabled you can often catch the model doing something else while it waits for a command to finish for example. Or spin up multiple ones to run multiple test suites concurrently"  
[X Link](https://x.com/thsottiaux/status/2005157477177008494)  2025-12-28T06:02Z 18.8K followers, 11.1K engagements


"@12g8ge We use the CLI to experiment a bit more and then bring the best features to the IDE extension. If you enable it using /experimental in the CLI then it will take effect everywhere though and you don't miss out here"  
[X Link](https://x.com/thsottiaux/status/2005164210901450967)  2025-12-28T06:29Z 17.8K followers, [----] engagements


"@anayatkhan09 Not explicitly but you can let Codex know you have a small machine and to be careful and it will adapt the strategy. Or you can disable the feature when minimal resource usage is needed"  
[X Link](https://x.com/thsottiaux/status/2005166234959306842)  2025-12-28T06:37Z 17.8K followers, [----] engagements


"@thdxr With the models we have it's a bit ridiculous you can tell these apart so easily"  
[X Link](https://x.com/thsottiaux/status/2005167411851657221)  2025-12-28T06:42Z 17.7K followers, [----] engagements


"@webdevcody We need new tooling for open source maintainers. Part of having our codex repo open-source is that it forces us to think about the consequence of code writing becoming essentially free"  
[X Link](https://x.com/thsottiaux/status/2005167729272348673)  2025-12-28T06:43Z 17.8K followers, [----] engagements


"@lamps_apple We don't have it as a native feature but you can do pretty much anything through a shell. That means that codex can use codex or any other agent if they have a cli"  
[X Link](https://x.com/thsottiaux/status/2005168209926959385)  2025-12-28T06:45Z 17.8K followers, [----] engagements


"@nearcyan Codex is still early"  
[X Link](https://x.com/thsottiaux/status/2005169048787861901)  2025-12-28T06:48Z 17.8K followers, [----] engagements


"@adamdotdev Especially true when you have @steipete as a user"  
[X Link](https://x.com/thsottiaux/status/2005169552351715556)  2025-12-28T06:50Z 17.8K followers, [----] engagements


"With Codex we aim for simplicity. You dont need a complex setup it just works. You should be able to open it for the first time or the thousand time and not need to configure much or anything for it to work optimally. The one advice I do give is to ask more questions. my brothers in christ codex is like [--] times more cracked than claude code . without rules hooks plugins skills and a bunch of other bullshit the mf cooks for 30-60 mins and the thinking process is actually insane when u read it . but it's sloooooow af and not "vibey" enough my brothers in christ codex is like [--] times more"  
[X Link](https://x.com/thsottiaux/status/2005433644803137818)  2025-12-29T00:20Z 21.4K followers, 85.4K engagements


"@glawsontweets We will keep improving the DX while optimizing for simplicity. You should not need [--] knobs to unlock stellar performance"  
[X Link](https://x.com/thsottiaux/status/2005469071962472561)  2025-12-29T02:41Z 17.8K followers, [---] engagements


"@NektariosAI @OpenAI @sama Whats missing from running codex directly within powershell"  
[X Link](https://x.com/thsottiaux/status/2005493467582234754)  2025-12-29T04:17Z 17.8K followers, [----] engagements


"@timurkhakhalev Well ship an improvement for this soon its a point of friction right now"  
[X Link](https://x.com/thsottiaux/status/2005518747088674831)  2025-12-29T05:58Z 17.8K followers, [---] engagements


"@kr0der @steipete Let me know as you go what we can improve big or small. Hope you enjoy the ride"  
[X Link](https://x.com/thsottiaux/status/2005544642549559655)  2025-12-29T07:41Z 18.4K followers, [----] engagements


"Asked codex to go through my /.codex/sessions and extract avg and median time per turn for GPT-5.2-Codex at medium and xhigh reasoning and it shows that medium is 4-5X faster. This is why I recommend using medium as your default and using xhigh for background multitasking"  
[X Link](https://x.com/thsottiaux/status/2005564827360809232)  2025-12-29T09:01Z 22.4K followers, 42.8K engagements


"@defido For when you want to get something done to the highest degree of quality and don't mind waiting. Which also means that you are likely to go do something else such as scrolling twitter or engage in a different codex session"  
[X Link](https://x.com/thsottiaux/status/2005565512059871551)  2025-12-29T09:04Z 22.9K followers, [----] engagements


"@SystemSculpt If you are happy to juggle a few in progress tasks then nothing beats xhigh in terms of quality of results"  
[X Link](https://x.com/thsottiaux/status/2005565821830230092)  2025-12-29T09:05Z 17.8K followers, [----] engagements


"Coding agents are only constrained by physics. Laptops and the entire dev stack were built for humans. We are still very early"  
[X Link](https://x.com/thsottiaux/status/2005809050890686561)  2025-12-30T01:11Z 21.4K followers, 19.9K engagements


"@GlitchInTheCode Something in the same spirit yes"  
[X Link](https://x.com/thsottiaux/status/2005810566967730271)  2025-12-30T01:17Z 17.9K followers, [----] engagements


"@tginart @kr0der @steipete You can just codex login"  
[X Link](https://x.com/thsottiaux/status/2005963833689661883)  2025-12-30T11:27Z 17.9K followers, [---] engagements


"@steipete @mitsuhiko Alarm every [--] minutes works well"  
[X Link](https://x.com/thsottiaux/status/2005973790610887011)  2025-12-30T12:06Z 18.2K followers, [----] engagements


"Common shape of Codex trace on hard problems (using GPT-5.2-Codex). Round of context gathering followed by heavy thinking then edits then round of double checking the work and making small adjustments. Clear that a lot of the time is spent inferencing and not toolcalls"  
[X Link](https://x.com/thsottiaux/status/2005977876022321426)  2025-12-30T12:22Z 21.7K followers, 25K engagements


"@kr0der We're looking at ways to make it easier to track multiple concurrent codex threads. Will DM you when there is something to try"  
[X Link](https://x.com/thsottiaux/status/2006317201692016909)  2025-12-31T10:51Z 20.3K followers, [----] engagements


"@simonw Just aligning the documentation with how folks refer to it. I personally differentiate between cloud tasks and codex web. With cloud tasks running on our hosted runtime (includes code review github slack linear .) and codex web being the web app"  
[X Link](https://x.com/thsottiaux/status/2006421779246624875)  2025-12-31T17:46Z 18.2K followers, [----] engagements


"@trashh_dev Need to revise your friend circle"  
[X Link](https://x.com/thsottiaux/status/2006610477586206736)  2026-01-01T06:16Z 18.4K followers, [----] engagements


"@yacineMTB The flips. The flops. See you back in a few days"  
[X Link](https://x.com/thsottiaux/status/2006744297580933561)  2026-01-01T15:08Z 18.7K followers, 14.5K engagements


"@jarrodwatts Let me know if you encounter anything that can be improved (other than speed or more collaborative personality we're working on those). Excited to see what your experience is as some new to codex"  
[X Link](https://x.com/thsottiaux/status/2007123802493473081)  2026-01-02T16:16Z 18.5K followers, [----] engagements


"@YuriKushch @jarrodwatts Only displayed if you've setup your repo through this requires visiting it on desktop. Once setup you can schedule cloud tasks through the iOS app https://chatgpt.com/codex https://chatgpt.com/codex"  
[X Link](https://x.com/thsottiaux/status/2007132367316725908)  2026-01-02T16:50Z 18.8K followers, [---] engagements


"The way I feel is that curating your own skills under /.codex/skills is the closest you get to training your own pokmon"  
[X Link](https://x.com/thsottiaux/status/2007136192467497317)  2026-01-02T17:05Z 22.3K followers, 12.3K engagements


"@robinebers Both use this https://platform.openai.com/docs/guides/priority-processing https://platform.openai.com/docs/guides/priority-processing"  
[X Link](https://x.com/thsottiaux/status/2007368847758971350)  2026-01-03T08:30Z 18.6K followers, [----] engagements

Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing