https://chat.lmsys.org/

Resolves to:
1st: 50%

2nd: 30%

3rd: 20%

Based on ELO rankings, only top score for each lab counts.
With apologies the the open-source community, tuned models will resolve to the underlying base model.

For example, below (leaderboard as of Nov 2023) would resolve
OpenAI 50%, Anthropic 30%, Meta 20%