Anthropic launched two new AI models, Claude Opus 4 and Claude Sonnet 4. The company built special safety tools to prevent people from making dangerous weapons with these programs. Opus 4 works best for writing computer code and handling tough jobs, while Sonnet 4 helps developers with daily tasks. Both models think better and remember more than older versions.
Opus 4 scored high marks on coding tests and problem-solving challenges. GitHub will use Sonnet 4 to power its new helper tool for programmers. The models take fewer shortcuts when solving problems compared to earlier versions. They can run multiple tasks at the same time and keep track of files better. Anthropic added new features that let the AI run Python code and connect to outside programs.
The company activated its strongest safety rules for Opus 4. These rules watch what people ask the AI and block dangerous requests about weapons. Anthropic uses special filters that check conversations in real time. The company also set up more than 100 security steps to protect the AI from theft. Claude Code became available to all users and works with popular coding programs.
Opus 4 scored high marks on coding tests and problem-solving challenges. GitHub will use Sonnet 4 to power its new helper tool for programmers. The models take fewer shortcuts when solving problems compared to earlier versions. They can run multiple tasks at the same time and keep track of files better. Anthropic added new features that let the AI run Python code and connect to outside programs.
The company activated its strongest safety rules for Opus 4. These rules watch what people ask the AI and block dangerous requests about weapons. Anthropic uses special filters that check conversations in real time. The company also set up more than 100 security steps to protect the AI from theft. Claude Code became available to all users and works with popular coding programs.
