Give every model the same prompt and compare the answer, the generated code, a live preview, speed, tokens and cost — all streaming in real time, right in your browser.