Apple Partners With Anthropic for Claude-Powered AI Coding Platform

Friday May 2, 2025 11:10 am PDT by Juli Clover

Apple is working with Anthropic on an updated version of Xcode that will support AI code writing, editing, and testing, reports Bloomberg. Anthropic is best known for its "Claude" large language model and chatbot that competes with OpenAI's ChatGPT. Claude is well-known for its coding capabilities, beating out other LLMs on programming tasks.

iu 2 1
The new version of Xcode integrates the Claude Sonnet model, and Apple is slowly rolling it out internally for employees to use. Many companies have started adopting AI coding tools to speed up product development, and Apple can't afford to be behind. Bloomberg says that Apple was initially hesitant to use AI to create software for Apple customers, but advancements in LLMs have made the company more comfortable with expanded AI tools.

Apple's internal version of Xcode has a chat interface for programmers to request code or help with coding, and it is able to test user interfaces and help engineers locate and address bugs.

There's no word yet on whether Apple will eventually launch the software publicly, but if the company's internal testing is successful, it's possible the tool could be provided to third-party developers. Last year, Apple announced Swift Assist, an AI-powered coding companion for Xcode. Swift Assist was supposed to come out in 2024, but like the personalized Apple Intelligence Siri features, it never materialized.

As it turns out, Apple engineers testing Swift Assist found that it was making up information and could slow down app development in some situations. Apple's work with Anthropic could help with Swift Assist and future tools.

Apple has inked deals with several AI companies. OpenAI's ChatGPT is already integrated into iOS 18, it is working with Google to add support for Gemini, and now there is a partnership with Anthropic. In China, Apple is also working with Baidu and Alibaba on AI.

Tag: Anthropic

Popular Stories

Anthropic Says Users Can't Be Needlessly Cruel to Claude

Thursday October 8, 2026 2:53 pm PDT by Juli Clover

Anthropic today updated its usage policy (via The Verge) to bar users from subjecting Claude to "sustained and needless abusive or cruel behavior." Anthropic says the abuse rule is meant to apply only in extreme cases, where users "repeatedly act cruelly" toward the AI models with "no discernible purpose." The rule does not apply to ordinary user frustration, pushback, dark creative themes,...

• 128 comments

Anthropic Launches Claude Opus 5.5 With Fable-Level Performance at a Lower Price

Tuesday September 22, 2026 12:23 pm PDT by Juli Clover

Anthropic today launched Claude Opus 5.5, which it says is the first model in a new 5.5 model family. Opus 5.5 performs at the level of Claude Fable 5.1 on most work, but costs 40 percent less to run than Opus 5. According to Anthropic, Opus 5.5 is a "major step up" from Opus 5 with improvements in performance and safety. Opus 5.5 is comparable to Mythos 5.1 in biology and cybersecurity, so...

• 38 comments

Three Years of 80% Charge Limits on iPhone: The Results

Tuesday October 6, 2026 12:59 pm PDT by Juli Clover

Since 2023, Apple's newer iPhones have had an 80 percent charge limit for those who want to extend battery lifespan. Charging to 80 percent instead of 100 percent is supposed to put less stress on the battery, improving longevity. For the last three years, I've used that setting on my iPhone. I started with the iPhone 15 Pro Max, then did the iPhone 16 Pro Max, and most recently, the iPhone...

• 418 comments

Top Rated Comments

novagamer

19 months ago

Sensible. Finally a good decision.

Score: 13 Votes (Like | Disagree)

Enlightened Doggo

19 months ago

We all have claude licenses at work. I've never once gotten code out of it that actually works with any of our codebases and neither have my coworkers. It will barf out 5x the code that is needed and it will just be wrong. People have setup entire projects for it adding all of the context needed and no success with even the most basic of tasks. It's so plainly unintelligent I would go as far as saying you shouldn't trust engineers who claim that it is.

Score: 11 Votes (Like | Disagree)

Eric_WVGG

19 months ago

Do you think we’re closer to that “vibe coding” paradigm where we will be able to code our own software without coding knowledge? I know it’s still not possible, but I am convinced that sooner or later, that day will come.

The problem with vibe coding is that it's just an elaborate way of copy-pasting work that already exists, and leading to a new generation of programmers who don't know how to write new code in the first place. Taken to its logical extreme, tech will stagnate, and nobody will know how anything works anymore.

But I'm not really worried about it… I think this vibe code **** is just a dumb fad, and a lot of kids are gonna find themselves fired.

Score: 10 Votes (Like | Disagree)

PsykX

19 months ago

As a software engineer, every line of code needs to meet my standards, including how variables are named.

No problem with that.

Wrong use of AI (95% people fall into this trap!) : Use AI to blindly write code, copy-paste it in your dev tool, test it a little bit, and release to public if it seems to work
Good use of AI : Use AI to write code, ask it to follow a specific architecture, review code thoroughly, improve it, use AI to comment the code, use AI to create unit tests for its own code, etc.
Another good use of AI : You need a very small piece of code, and you know exactly what you need, but it will take you 10-15 min to write. Use AI, save yourself a few minutes, and you'll quickly see if the code makes sense or not.

If you don't adopt AI as a software dev, you'll quickly become less productive than another software dev.

Score: 9 Votes (Like | Disagree)

mozumder

19 months ago

While I've grown more trusting of AI chat as an alternative to a Web Search, there's no way that I'm letting AI write my code. As a software engineer, every line of code needs to meet my standards, including how variables are named.

Or maybe it learns "my style" and would generate code that's compliant? I'll need to watch that space closely.

As for Xcode, using AI may be a solution to me creating Mac/iPhone apps, as I've never quite crossed into that space having to learn a new language. This might bridge the gap.

It writes better code than you can.

Score: 9 Votes (Like | Disagree)