Anthropic’s New Claude AI Models and Computer Control: A Leap in AI Capabilities Anthropic has launched the latest iteration of its Claude AI models, Claude 3.5 Sonnet and Claude 3.5 Haiku, bringing groundbreaking advancements in AI-driven tasks, particularly in coding and human-like computer control. Claude 3.5 Sonnet has set new benchmarks, outperforming other models with a remarkable 49.0% on the SWE-bench Verified, excelling in coding tasks. Its computer control feature allows the AI to manipulate screens, click, and type just like a human. Meanwhile, Claude 3.5 Haiku offers cost-effective and speedy performance, on par with its predecessor Claude 3 Opus. This development marks a pioneering step toward AI systems that interact with real-world computer environments. Safety and Future Impact Anthropic emphasizes safety, with both the US and UK AI Safety Institutes involved in rigorous pre-deployment testing. Their Responsible Scaling Polic