AI is writing for us, testing for us… all that’s left is to fire us.
Apple has begun testing a new AI-powered coding platform that writes, edits, and tests code on behalf of developers. The system is being developed in collaboration with the startup Anthropic, whose Claude Sonnet model is already recognized as one of the best for code generation. According to sources, the new platform is embedded in a special version of Xcode, Apple’s primary development environment. For now, it's being used internally, and there's no decision yet on a public release.
Previously, Apple avoided using generative AI in programming, but the rapid rise of large language models has changed its stance. A year ago, the company introduced Swift Assist, an in-house AI tool for Xcode, but it never made it to public release. Internally, engineers criticized it for being inaccurate and slowing down workflows. The partnership with Anthropic is a sign that Apple now recognizes the value of external AI solutions. Both Swift Assist and the new Claude-powered tool are expected to coexist.
The Claude Sonnet-based tool includes a chat interface where developers can ask questions and receive code or edits. It can also test user interfaces and assist with debugging — time-consuming tasks for many programmers. Notably, Apple has typically avoided integrating third-party AI models, with the exception of its deal with OpenAI, which brought ChatGPT support to Siri. Google’s Gemini model is also expected to be integrated in the future.
This platform could be a game-changer if Apple eventually opens it up to third-party developers — especially since Xcode and Swift are the main tools for building apps for iPhones and other Apple devices. More details are likely to be announced at Apple’s developer conference, which begins June 9.
For Anthropic, this collaboration is a major milestone. While the startup is already working with Amazon to enhance the Alexa+ voice assistant, it didn’t make Apple’s final shortlist last year when the company was seeking AI partners for Siri. That decision came down to OpenAI and Google — and ChatGPT ultimately won.
As AI-powered development tools like GitHub Copilot and Windsurf gain traction, Apple is trying to catch up. Its own AI efforts have faced criticism for lagging behind competitors. The company’s attempt to revamp Siri last year was also shelved indefinitely.
Apple is now undergoing internal changes. Some AI responsibilities have shifted from AI chief John Giannandrea to software head Craig Federighi, whose team now oversees Siri and several other AI initiatives. This marks a broader strategic shift toward a more decentralized, flexible approach to AI development.
During Apple’s Q2 earnings call, CEO Tim Cook confirmed that the company plans to combine in-house AI models with third-party solutions. He also emphasized Apple’s commitment to on-device AI processing, minimizing reliance on cloud data. According to Cook, progress is underway — and exciting innovations are just around the corner.
