Feb 15
Add setting for enabling thinking and setting thinking level
Hello team,Requesting you to kindly add a feature to add LLM level settings for enabling thinking mode and further mode, ability to add the level of thinking wherever applicable. For example, Gemini 3 models have minimal, low, medium and high depending on whch model you are using. Similarly, GPT 5.2 has thinking mode as well as deep research. In the same line, Anthropic models have ability to toggle Extended thinking mode. It would be great if we can add these toggles to the specific LLMs which support it so that we can get the best result out of all of these LLMs as per our task requirements.
Pending
YES! I can't believe this still hasn't been implemented. PLEASE add the option to manually set the reasoning effort parameter especially in the regular chat mode (Multi AI Chat). I understand 1minAI wants to make the UX easy to use to conform to a wider audience and adding advanced parameter control might go against that. However unlike specific stuff (like Temperature parameter and repeat/frequency penalty), I believe the ability to set the reasoning effort is an INTEGRAL feature due to its immense usefulness (and can't be set for sure just by telling the model to do so since it is low level). As OP mentioned, due to the variety of reasoning parameter options (for ex. GPT 5.6 Sol's wide variety: low/medium/high/xhigh/max), the easiest (for devs) way to implement this would be a reasoning effort string that the user manually types in but the most convenient way for the user is to implement it as switches/sliders for each models (each models' documentation indicate which parameters are supported so it's not a difficult info to find).