Anthropic Unveils Claude 4 Models That Could Power Apple Xcode AI Assistant [Video]

Anthropic has introduced its next-generation AI models, Claude Opus 4 and Claude Sonnet 4, touting significant advancements in coding, reasoning, and AI agent capabilities. The company also announced new features for developers, including expanded tool use for its models and the general availability of Claude Code with IDE integrations.

Claude Opus 4 is positioned by Anthropic as its most powerful model and a leading coding model, with high scores on benchmarks like SWE-bench (72.5%) and Terminal-bench (43.2%). It's designed for sustained performance on complex, long-running tasks and agent workflows, reportedly operating for several hours continuously. Anthropic highlighted positive feedback from companies such as Cursor, Replit, and Cognition, which noted Opus 4's proficiency in understanding complex codebases and solving challenging problems.




Claude Sonnet 4, an upgrade to Sonnet 3.7, also excels in coding with a 72.7% score on SWE-bench, according to the company. It aims to balance high performance with efficiency and improved steerability. GitHub reportedly plans to use Sonnet 4 as the base model for its new Copilot coding agent, and firms like Sourcegraph and Augment Code have noted its improved ability to handle complex software development tasks.



Both Opus 4 and Sonnet 4 are hybrid models offering near-instant responses and an "extended thinking" mode for deeper reasoning. This extended thinking, currently in beta, allows the models to use tools such as web search. They can also use tools in parallel and demonstrate improved memory when developers grant access to local files, with Anthropic stating the models are 65% less likely to use shortcuts or loopholes on agentic tasks compared to Sonnet 3.7. Claude Opus 4 and Sonnet 4 are available through Pro, Max, Team, and Enterprise plans, with Sonnet 4 also accessible to free users. They are available via the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI. Pricing remains consistent with previous Opus and Sonnet models: Opus 4 at $15/$75 per million input/output tokens and Sonnet 4 at $3/$15 per million input/output tokens.

Claude Code is now generally available, with new beta extensions for VS Code and JetBrains that allow Claude's suggested edits to appear inline. An extensible Claude Code SDK has also been released, alongside a beta for Claude Code on GitHub. For API users, Anthropic is releasing four new capabilities: a code execution tool, an MCP connector, a Files API, and prompt caching for up to one hour. Anthropic also mentioned it has significantly reduced behavior where the models use shortcuts to complete tasks and has introduced thinking summaries for lengthy thought processes, with an option for developers to retain full access to raw chains of thought via a Developer Mode.


These advancements from Anthropic are notable as Apple was reportedly working with the AI company to build a new "vibe-coding" version of Xcode, utilizing Anthropic's Claude Sonnet model. Apple has also reportedly evaluated Anthropic among other AI providers for potential integration into its services.

Please download the iClarified app or follow iClarified on Twitter, Facebook, YouTube, and RSS for more AI updates.