Suggestions regarding "thinking" models after v1.5.40

[LFd3v]

c/c: @HectorEspejo

First of all, thank you a lot for the PR, it seems to work fine, finally it is possible to actually set the reasoning effort for GPT-OSS.

Some suggestions:

1. Remove "Thinking model (Ollama)" toggle from Current Chat Model Settings as well as from Model Settings in Settings -> Manage models, since it is not needed anymore, right?
2. Do the same for "Reasoning Effort" text entry;
3. As mentioned in another comment Ollama API can be used to check if a model has "thinking" capabilities. This is important, not only because this allows the toggle to be used with more models but also because filtering by name can result in many problems because Ollama returns an error if the mode is enabled for a model that actually does not support it (not all Qwen3 models supports it, for instance, Qwen3-Coder doesn't).

Thanks again. Regards

[n4ze3m]

Yep, really good ideas