LLM Leaderboardss

[Kreijstal]

not sure if it still makes sense but we could still test quants, or local models, and update the benchmarks, also still curious how Fable 5 would score

[dwash96]

I feel like the polyglot benchmark is getting pretty saturated but I would definitely publish local model benchmarks and if there is anything on openrouter you think we should run, I can set up those runs later this week for that set