Anthropics neuer Claude Opus 4 kann sieben Stunden lang autonom laufen

„On Thursday, Anthropic announced Claude Opus 4 and Claude Sonnet 4, its next generation of models, with an emphasis on coding, reasoning, and agentic capabilities. According to Rakuten, which got early access to the model, Claude Opus 4 ran „independently for seven hours with sustained performance.“

Alongside the launch of Opus 4 and Sonnet 4, Anthropic also introduced new features. That includes web search while Claude is in [extended thinking](https://www.anthropic.com/news/visible-extended-thinking) mode, and summaries of Claude’s reasoning log „instead of Claude’s raw thought process.“ 

In the safety and alignment realm, Anthropic said both models are „65 percent less likely to engage in reward hacking than Claude Sonnet 3.7.“ Reward hacking is a slightly terrifying phenomenon where models can essentially cheat and lie to earn a reward (successfully perform a task).“

We’ve invented a technology that could bring us into the future and make the world a better place for everyone, and solely because of capitalist greed, we’re instead using it to speed-run both oligarchic dystopia and planetary environmental annihilation.