# Challenges With Codex - Comparison with GitHub Copilot and Cursor **URL:** \ **Category:** Codex\ **Created:** [September 18, 2025, 11:14am UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767 "2025-09-18T11:14:49Z")\ **Posts on this page:** 11\ **Page:** 1
**Author:** ![gmoore](https://avatars.discourse-cdn.com/v4/letter/g/47e85d/32.png) [@gmoore](https://community.openai.com/u/gmoore)\ **Post date:** [September 18, 2025, 11:14am UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/1 "2025-09-18T11:14:49Z")
I’m having some trouble developing code with Codex. I am using Codex in VS Code Insiders. The primary issue is Codex tends to insert /r/n characters throughout code files. Then the code will not build. Is there a way to get it to stop doing this? Also, I think there needs to be a middle-ground between Agent mode and Agent (full access). With agent mode, I am prompted for every single command to run. This would have me clicking “Approve” all day long. For Agent (full access), it’s almost too much access. I need a solution that is somewhere in between the two. Finally, no matter which Codex model I pick, the editing capability seems to be about 10% as efficient as GitHub Copilot or Cursor using the premium models within those IDEs. I want to be able to use Codex, but each time I give it a shot it is inefficiently making edits that are off and don’t even compile. For example, I waited 30 minutes for Codex to struggle through making edits that didn’t compile - malformed html tags, erroneous \r\n characters. GitHub Copilot and Cursor are able to make the same changes correctly in 10% of the time that Codex is taking on low, medium or high models. At this point, Codex isn’t usable. Please help. ---
**Author:** ![Jack\_Devoy](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/jack_devoy/32/681426_2.png) [@Jack\_Devoy](https://community.openai.com/u/Jack_Devoy)\ **Post date:** [September 19, 2025, 7:07am UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/2 "2025-09-19T07:07:41Z")
I am having same issue. It likely an issue with the powershell commands, I’m not sure what it’s like on mac os but i’m having issue on windows 11. And then it was putting like alien characters in code like ‘aaaaAAaaaaAaaAaaAA” but different variants of “a” or something. ---
**Author:** ![gmoore](https://avatars.discourse-cdn.com/v4/letter/g/47e85d/32.png) [@gmoore](https://community.openai.com/u/gmoore)\ **Post date:** [September 19, 2025, 11:57am UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/3 "2025-09-19T11:57:53Z")
Which IDE are you using @Jack_Devoy ? I am also on Windows 11. ---
**Author:** ![Jack\_Devoy](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/jack_devoy/32/681426_2.png) [@Jack\_Devoy](https://community.openai.com/u/Jack_Devoy)\ **Post date:** [September 19, 2025, 12:11pm UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/4 "2025-09-19T12:11:53Z")
I am using cursor. With codex extension. i’ve also tested in terminal and i think there was thge same issue. I am usign vscode for another project atm but i don’t think i’ve seen it yet but i’m still building out react nativie app. So it may be due to it searching through large codebase and a tool call it uses to search and edit. It could also be how cursor ide wraps vscode but i can’t confirm if issue is just on cursor yet. ---
**Author:** ![gmoore](https://avatars.discourse-cdn.com/v4/letter/g/47e85d/32.png) [@gmoore](https://community.openai.com/u/gmoore)\ **Post date:** [September 19, 2025, 12:23pm UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/5 "2025-09-19T12:23:52Z")
I’m having the issues in VS Code, so it isn’t only Cursor. ---
**Author:** ![MichiStein](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/michistein/32/423206_2.png) [@MichiStein](https://community.openai.com/u/MichiStein)\ **Post date:** [September 19, 2025, 12:51pm UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/6 "2025-09-19T12:51:07Z")
I am on Windows 10 with Cursor, and usually the errors lie also in the Power Shell Terminal Commands, although not quite as often as you described ---
**Author:** ![Mr\_Nice\_Guy](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/mr_nice_guy/32/682678_2.png) [@Mr\_Nice\_Guy](https://community.openai.com/u/Mr_Nice_Guy)\ **Post date:** [September 23, 2025, 1:37am UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/7 "2025-09-23T01:37:44Z")
Did you say, Windows….. Of course you mean Linux right? No one would willfully subject themselves to Windows. I assume that your Linux machine is temporarily broken, and you’ve been forced to borrow a non-developer friends computer? I had a lot of similar problems with it running powershell commands that all went away when I took Linux’s hand and sailed off into the developer sunset. ---
**Author:** ![peter.wyss](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/peter.wyss/32/511302_2.png) [@peter.wyss](https://community.openai.com/u/peter.wyss)\ **Post date:** [September 26, 2025, 9:21am UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/8 "2025-09-26T09:21:28Z")
* * * Same here. My experience so far comparing **Visual Studio Code with the Codex Extension** vs. **Visual Studio Code with GitHub Copilot** : - **Codex** is noticeably slower. - There’s no way to preview or track changes made by the agent (no Undo/Apply workflow like in Copilot), which makes it hard to see what actually changed between prompts. → Edited: This seems to be working now. - I also ran into issues with `\r\n` characters (as mentioned earlier) in Codex. - I tested with **gpt-5-codex-medium**. Overall, given these issues, it doesn’t feel usable in its current state. Would be great to see some improvements in the near future. Now that we can also use the new Codex model within GitHub Copilot: > **[OpenAI GPT-5-Codex is rolling out in public preview for GitHub Copilot -...](https://github.blog/changelog/2025-09-23-openai-gpt-5-codex-is-rolling-out-in-public-preview-for-github-copilot/)** > > OpenAI’s GPT-5-Codex model, optimized for agentic coding, is now rolling out in GitHub Copilot Availability in GitHub Copilot GPT-5-Codex will be available to Copilot Pro, Pro+, Business, and Enterprise. You’ll… Has anyone tried **Visual Studio Code + GitHub Copilot (with gpt-5-codex)** and compared it directly to using Codex itself (same model)?strong text ---
**Author:** ![gmoore](https://avatars.discourse-cdn.com/v4/letter/g/47e85d/32.png) [@gmoore](https://community.openai.com/u/gmoore)\ **Post date:** [September 26, 2025, 10:14am UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/9 "2025-09-26T10:14:19Z")
Yes GPT-5-codex in GitHub Copilot or Cursor works quite well. Give that a try. ---
**Author:** ![keesan](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/keesan/32/740295_2.png) [@keesan](https://community.openai.com/u/keesan)\ **Post date:** [June 8, 2026, 5:06am UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/10 "2026-06-08T05:06:17Z")
A lot of the difference people feel here is not raw model quality, it’s whether the runtime knows when to stop. The moment a loop can explain what failed, what changed, and why the next attempt is better, the whole experience feels more trustworthy. Without that, the system can look smart right up until it gets expensive. ---
**Author:** ![Prashant\_Pardesi](https://avatars.discourse-cdn.com/v4/letter/p/f9ae1b/32.png) [@Prashant\_Pardesi](https://community.openai.com/u/Prashant_Pardesi)\ **Post date:** [June 17, 2026, 3:44pm UTC](https://community.openai.com/t/challenges-with-codex-comparison-with-github-copilot-and-cursor/1358767/11 "2026-06-17T15:44:49Z")
Closed on June 17, 2026, 3:44pm UTC