Hello! Often when conversing with models (e.g., Grok), I'll turn up the speed so they talk faster, usually something like 1.5x but sometimes up to 2x. It would be great if there was an env var or something that I could configure this with. Thanks!