when will we be able to select GPT5 for our assistant models?

I recently watched a customer interact with a wudget button on a website. it was a chrome browser on a mac. when he clicked the button, the assistant greeted him, then heard herself and started going back and forth answering itself... I tried lowering the volume on the mac , even the mic, but that didnt fix the issue.... so, thoughts? anyone experience a similar sitution??

Any chance you guys can optimize/reduce the JS on your side? ------------------------------------------------------------------------------------------------------------- 🎯 Here’s the important thing: Even though you lazy-loaded the iframe,GTMetrix still sees the iframe's payload because: ➡️ GTMetrix "scrolls" the page automatically when it tests (to simulate real user behavior).➡️ When it scrolls, the iframe loads, and those external scripts still download.➡️ Therefore, the heavy scripts still hurt your LCP (Largest Contentful Paint). 🔥 So what can you realistically do now? You have two paths: 1. Accept some weight if the iframe is important - You already minimized the impact by lazy loading. - You can't control the size of createassistants.com scripts — those are external (third-party hosted). - GTMetrix will always count third-party scripts if they get triggered at any point. ✅ This is acceptable if the assistant is important for UX and conversions. 2. Delay iframe even further (optional next level optimization) We can make it even more "deferred": - Instead of loading when visible - Load the iframe only after 5–7 seconds (even if the user never scrolls). This way: - GTMetrix will not catch it unless it sits there for 7 seconds. - Improves LCP / total payload in GTMetrix. ⚡ Example:Add a setTimeout to delay injecting the iframe. Would you like me to show you this version too? It’s super easy. 📈 How bad are those scripts really? Let's zoom in a bit: URLSizecreateassistants.com/run.js 801 KB createassistants.com/static.js 738 KB createassistants.com/dynamic.js 592 KB Total for those 32.13 MB! 🚨 That's more than half your total page weight! 🧠 Bottom Line: - You've already done everything you can on your side to optimize the iframe loading behavior. - The size of external scripts cannot be minimized by you, unless the service owner compresses their files.

looking at some call activities, the orbs don't stop recording even after the orb conversaton ended. I am seeing multiple 30minute calls just recording converations that are not addresed to the orb and sometimes people here 2 orbs talking. can we set a maximum time for the call and have the orb de-activated?