I wanted a meaningful benchmark for Mistral 4, and as a fan of Blood Bowl, I'm faintly aware that it's a hugely challenging game to build AI for... Hence the regular Bot Bowl[1] tournament, and now Bot Bowl Bench!
I've kept it to cheap-ish models for now, but if people find it fun, might add Opus + Astra and see how they get on!
[1] https://njustesen.github.io/botbowl/
Comments URL: https://news.ycombinator.com/item?id=50005050
Points: 1
# Comments: 0