New emergency update to ReaLackey, v0.3.1, fixes issues when trying to use GPT5+ models win ReaLackey. Also the default model for Gemini defaults to 3.5-flash now. https://github.com/Timtam/ReaLackey/releases/tag/v0.3.1