Model picker in the LeChat App (and persistent reasoning)
I am trying to migrate from Claude to LeChat. I am paying for Pro. I do not get why there is no model picker in the LeChat app. For every prompt I am wondering if this request was just executed by an uncapable outdated model. The answers are really inconsistent, sometimes they are really good, sometimes they are trash. Having no control over the model that is being used is just frustrating and reduces trust in the answers, especially when replies are bad. And it is really annoying that I do not get to persistently activate thinking. Everytime I open the app, i have to reconfigure thinking / reasoning before sending a prompt. There is a reason every single other chat app gives users control over the model, especially paying users. With the claude pro plan, i get to use opus in the chat with almost never running into any limits, and mistral medium 3.5 is more than 3 times cheaper on the api, yet i am not able to consitently use it, while the subscription pricing is roughly the same (or at least, i dont know if it is used, i hope not).
I really want to support mistral, but with the inconcistency in replies, which I hope are caused by worse models being used, I cant justify using mistral over claude. I can still use mistral for super simple queries, but for anything that gets a bit complicated, i dont feel confident using lechat as i have no idea how competent the model being used is. A model picker would really change.