1import"../chunks/CWj6FrbW.js";import"../chunks/69_IOA4Y.js";import{Y as a,at as o,b0 as n}from"../chunks/CBwW-_qi.js";var s=o('<div class="bg-gray-100 rounded-lg md:rounded-[15px] p-6 md:p-8 flex-1 not-prose"><p class="text-base md:text-lg">Can models click UI elements on a desktop computer?</p></div> <p>In this benchmark, LLMs are tasked with analyzing a computer screenshot to predict the pixel coordinates of a specified UI element. The UI element is described using natural language instructions â for example, âclick the Get in Touch buttonâ.</p> <p>Performance is evaluated based on two criteria: first, the frequency with which the predicted coordinates fall within the bounding box of the target UI element; and second, the distance from the predicted coordinates to the nearest edge of the bounding box when predictions fall outside it.</p> <p><img src="/images/datasets/computer-use/model_hit_accuracy.png" alt="Computer Use Chart 1"/></p> <p><img src="/images/datasets/computer-use/model_distance.png" alt="Computer Use Chart 2"/></p> <div class="bg-gray-100 rounded-lg md:rounded-[15px] p-6 md:p-8 flex-1 not-prose"><h3 class="text-xl mb-4">Custom for your agent</h3> <p class="text-base md:text-lg">We believe in the future of general-purpose agents, with computer use being a significant part of that vision. We are actively working on developing more complex multi-step benchmarks. If you need to evaluate your agent, contact us at <a href="mailto:[email protected]" class="underline">[email protected]</a>.</p></div>',1);function p(e){var t=s();n(10),a(e,t)}export{p as component};
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.