Hi there and thanks for bringing this Gem to AS!

Offering more variety in the available LLMs would give users greater flexibility and control over how summaries are generated. While you currently support GPT-4o and Claude which are decent ones, expanding the range of powerful yet cost effective models — such as DeepSeek V3, Qwen Turbo, Mistral, Gemini, or Meta's Llama — would provide a richer experience and allow for more tailored results depending on user preferences.

I have updated to T3 hoping to hear positive outcome on this or any sort of OpenRouter Integration in the near future. This will democratize our summaries with the LLM we are more conformable with (you still can recommend GPT-4o or Claude as default user selection).

Thanks again!

Best,