I use some third-party providers for some of my AI models, with one of the latest additions being DeepSeek for their reasoner model. However, one of my providers has the model labeled as "TA/deepseek-ai/DeepSeek-R1" instead of the usual "deepseek-reasoner." I think that's why the "thinking" labels are not working and the usual thinking box is not showing (see image attached).

Is there a way to add, inside the customer model settings, a checkbox to "force" the recognition of these thinking labels so that they are not dependent on a specific model name? There seem to be more providers labeling or even mislabeling their models in a different way.