Inquiry Regarding Pairwise Model Comparison in Multi-Modality Arena

[zhimin-z]

Thank you for your remarkable contributions!

I've explored the multi-modality arena and noticed that it actually differs from the Chatbot Arena, where two anonymous models are compared side-by-side.

After playing with the demo in the README (as shown above), I observed that only one model is provided for evaluation by the third-party crowd:

I cannot find any arena-related keywords in the demo as well:

This leads me to inquire: where can we find the second model for conducting a pairwise comparison?
@shepnerd @wqshao126 @zzyfd @orashi