Summerize and compare all results

[PhillipRt]

Is your feature request related to a problem? Please describe.
If you are running 5-10 models, it comes very difficult to read though all of them, especially if they are contradicting
Describe the solution you'd like
Chatall wrote a promt that gets leads to a ranking of the answers and a final conclusion. Something similar would be nice.

Additional context
I know scraping the response form the different UIs is difficult but i beleve its worth it. A project called smart gpt tied this approach and beat a bunch of benchmarks with it

[PhillipRt]

#132 already touched on smartgpt but i think the chatall implementation still solves the 'no api keys' requirement

[swyxio]

cool, would happily welcome a PR for that!