Journal · May 5, 2026
Claude 4.5 Released: Advanced Visual Reasoning and System Control
Anthropic's Claude 4.5 redefines visual AI. Discover its advanced reasoning, parsing complex charts, and executing system operations directly. Get insights into
Anthropic has launched Claude 4.5, setting new standards for visual analysis and system automation. The model features enhanced capabilities to read layout coordinates, parse complex charts, and execute terminal operations directly in sandbox environments.
UI Auditing and Code Generation
Claude 4.5's visual capabilities let it review webpage screenshots and detect accessibility violations, visual bugs, or alignment errors. It can then draft pull requests to resolve these issues, functioning as an automated QA engineer.
Agentic Execution Power
The model can run terminal commands, execute scripts, and verify its output locally. This allows it to debug build errors and run automated tests in staging environments autonomously. It can interact with developer environments, execute commands, and run tests until the code builds correctly.
Claude 4.5 is available via the Anthropic API and console starting today, with pricing set at $15 per million input tokens and $75 per million output tokens, supporting next-gen software automation.