[Enhancement] Add Gemini Flash-Lite 3.1

[Akila3]

1. Please add the Gemini Flash-Lite v3.1 model as a selectable option in the AI Translation function.
   Alternatively, adding a single entry called “Flash-Lite-Latest”.
   That always points to the newest Gemini Flash-Lite model would be even more convenient and future-proof.

2. Add advanced settings, to allow power users to manually configure key generation parameters: Temperature, Top-P (nucleus sampling) and the reasoning/thinking token budget (e.g., 512, 1024, 2048, etc.). This would give power users greater control and flexibility, which is very needed when there is no surrounding context being submitted to the AI.

[silentbil]

Thanks @Akila3 i will look into

[Akila3]

Any update?

[mvanhorn]

Implemented in #363.

[silentbil]

@mvanhorn see my comments in your PR.
u added 2.5 flash lite instead of 3.1