Anthropic released Claude Opus 4 and Claude Sonnet 4, calling Opus 4 the world's best coding model and emphasizing its ability to sustain long, complex tasks. The models supported extended thinking interleaved with tool use (beta), parallel tool execution, and memory when given local file access, and Anthropic reported a 65% reduction in shortcut-taking versus Sonnet 3.7. Opus 4 scored 72.5% on SWE-bench and 43.2% on Terminal-bench; Sonnet 4 scored 72.7% on SWE-bench. Claude Code moved from research preview to general availability with VS Code and JetBrains integrations, GitHub Actions support, and an SDK for building custom agents. Anthropic said the release came with measures for higher AI Safety Levels including ASL-3.