Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Apple Releases Xcode 26 Beta 7 With GPT-5 Support and Claude Integration

MacRumors

MacRumors

macrumors bot
Original poster
Apr 12, 2001
69,007
40,020


Apple today released the latest version of Xcode 26, introducing several new features that developers can take advantage of.

apple-developer-app-feature.jpg

ChatGPT in Xcode now supports GPT-5, so developers can choose to start new conversations with GPT-5 or GPT-4.1. GPT-5 is the new default option, with two models to select from. GPT-5 is optimized for quick, high-quality results that Apple says will work for most coding tasks.

For more difficult tasks, developers can select GPT-5 (Reasoning), which spends more time thinking before providing a response.

Claude in Xcode is now available in the Intelligence settings panel in the Xcode app, so users can add their existing paid Claude account and use Claude Sonnet 4 for coding tasks.

When Xcode 26 officially launches, the app will support coding help from either ChatGPT or Claude, providing users with the option to use their favorite chatbot. Developers are also able to bring API keys from other providers or run local models on Apple silicon Macs.

Article Link: Apple Releases Xcode 26 Beta 7 With GPT-5 Support and Claude Integration
 
Article Link
https://www.macrumors.com/2025/08/28/xcode-gpt-5-claude-integration/
  • Like
Reactions: sos47 and mhnd
They should let you switch on the fly with a console command, opening prefs to toggle back and forth is a pain. Important if you run out of usage and want to continue with another model for a few hours until it resets.
 
I've been using Claude opus via API in earlier Xcode 26 betas, its not bad for somethings but really struggles with maintaining patterns with complexity through multiple files, it also over biases bug fixes and kind of throws away the original goal just to fix something.

I've gotten into a habit of using LLMs to validate things, reduce boiler plate, write tests, etc.

I use Claude code in the project folder to find things, make lists, etc. I have colleagues who submit whole PRs using AI tools and to date have resulted in a lot of re-work, from things like using print statements instead of using a logger, or adding singletons, ignoring existing dependencies, force unwraps and the comment puke also adds so much nonsense to what should nowadays be simple code, its kind of annoying.

I had high hopes for AI coding after the early GPTs but its cooled off a bit, I use it all day on one shot tasks and its great but it really falls apart quick.
 
  • Like
  • Love
Reactions: Vulkan, sos47, businesscats and 5 others
AppleTO said:
At what point does ChatGPT become the author of the app?

... kidding... sort of
Click to expand...
It's a valid question. If you get ChatGPT or Claude to generate the code, could there be any valid disputes about who really owns the IP?

It should be obvious that a tool is not a creator, but if the code itself hasn't been written by the developer....

... even messier, if the code turns out to be suspiciously similar to some non opensource code that the AI has "magically found".....

... it's this second part I'd worry about: "AI generated code" that turns out to be plagiarized code.

EDIT: I'm not being entirely serious either, but I wouldn't be surprised if a court case or two arise over the next months.
 
Last edited:
  • Like
Reactions: KeithBN, racerhomie, sos47 and 1 other person
MacRumors said:
GPT-5 is optimized for quick, high-quality results that Apple says will work for most coding tasks.

For more difficult tasks, developers can select GPT-5 (Reasoning),
Click to expand...
So ... GPT-5 is the default but also what you can switch to for more difficult tasks?
 
germanbeer007 said:
"hey bb, you wanna vibe code with me tonight?"
Click to expand...
There are 2 teams :

- Team AI

- Anti AI Team

---

Pros :

- AI work is faster than humans

- You don't need to know to code a lot, it can give you great results (sometimes)

- Can help coders to fix things, add things or do small things very fast

Cons :

- Not as good as a human work (In some cases)

- Models like chatgpt 5 thinking aren't free... if you use chatgpt 5 or 4.1, results can be bad.

- AI Pollute the earth

---

Local AI doesn't pollute and is free but local AI is not very intelligent...

In which team are you @everyone? (I'm in Team AI)
 
Last edited:
Astuces iOS said:
There is 2 teams :

- Team AI

- Anti AI Team

---

Pros :

- AI work is faster than humans

- You don't need to know to code a lot, it can give you great results (sometimes)

- Can help coders to fix things, add things or do small things very fast

Cons :

- Not as good as a human work (In some cases)

- Models like chatgpt 5 thinking aren't free... if you use chatgpt 5 or 4.1, results can be bad.

- AI Pollute the earth

---

Local AI doesn't pollute and is free but local AI is not very intelligent...

In which team are you @everyone? (I'm in Team AI)
Click to expand...

I'm in the "not doing this half-arsed and/or hallucinated crap" camp, otherwise known as the "anti-bubble" camp.

That'll be the AI camp if and when AI ever proves to be worth the billions that have been poured into it, but until then it's the other side of the table.

When AI becomes more of a tool instead of a crutch or distraction, ... but we're not at that point yet.

I don't get why you say Local AI doesn't pollute but server-side AI does. The only fundamental difference, in terms of resource consumption and so pollution, is scale.
 
  • Like
Reactions: KeithBN and sos47
dannyyankou said:
I’m not a developer, but I’m excited that these AI tools are coming to Xcode. It should hopefully make debugging apps more efficiency and timely.
Click to expand...
On the contrary, it actually stands to make debugging more difficult as the coder may have no idea why the code was implemented in the way it was. The method AI decides to use may not even work, so then you have a coder assuming the broken code should work and wondering what they’re doing wrong.
 
  • Like
Reactions: jw2002 and KeithBN
Happy_John said:
I'm in the "not doing this half-arsed and/or hallucinated crap" camp, otherwise known as the "anti-bubble" camp.

That'll be the AI camp if and when AI ever proves to be worth the billions that have been poured into it, but until then it's the other side of the table.

When AI becomes more of a tool instead of a crutch or distraction, ... but we're not at that point yet.

I don't get why you say Local AI doesn't pollute but server-side AI does. The only fundamental difference, in terms of resource consumption and so pollution, is scale.
Click to expand...
Yeah well when I mean local ai doesn't pollute well I mean it doesn't pollute a lot as others...
 
RMo said:
So ... GPT-5 is the default but also what you can switch to for more difficult tasks?
Click to expand...
There’s the baseline GPT-5 model and then another model with name suffixed (Reasoning) that’s better for more complex tasks. They’re both in the GPT-5 family of models.
 
  • Like
Reactions: SFjohn
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back