Polytopia Bench gives 12 models 40 minutes each to write a Polytopia bot in Rust from scratch.
The newly introduced benchmark asks how well frontier models can build bots for Polytopia. Its setup gives 12 models 40 minutes apiece to write a bot from scratch in Rust.
—
3 Sources, first seen ago
@kzdagoofHow good are frontier models at writing Polytopia bots? Introducing Polytopia Bench. 12 models each get 40 minutes to write a Rust bot from scratch.
@dwrRT @kzdagoof: How good are frontier models at writing Polytopia bots? Introducing Polytopia Bench. 12 models each get 40 minutes to writ…
@GregKamradtI’ve been waiting for this I’m not very good, but play for sport, mostly on airplanes Tiny map, dry land, 8 other players, domination@kzdagoofHow good are frontier models at writing Polytopia bots? Introducing Polytopia Bench. 12 models each get 40 minutes to write a Rust bot from scratch.@dwrRT @kzdagoof: How good are frontier models at writing Polytopia bots? Introducing Polytopia Bench. 12 models each get 40 minutes to writ…@GregKamradtI’ve been waiting for this I’m not very good, but play for sport, mostly on airplanes Tiny map, dry land, 8 other players, domination@kzdagoofHow good are frontier models at writing Polytopia bots? Introducing Polytopia Bench. 12 models each get 40 minutes to write a Rust bot from scratch.@dwrRT @kzdagoof: How good are frontier models at writing Polytopia bots? Introducing Polytopia Bench. 12 models each get 40 minutes to writ…@GregKamradtI’ve been waiting for this I’m not very good, but play for sport, mostly on airplanes Tiny map, dry land, 8 other players, domination@kzdagoofHow good are frontier models at writing Polytopia bots? Introducing Polytopia Bench. 12 models each get 40 minutes to write a Rust bot from scratch.@dwrRT @kzdagoof: How good are frontier models at writing Polytopia bots? Introducing Polytopia Bench. 12 models each get 40 minutes to writ…@GregKamradtI’ve been waiting for this I’m not very good, but play for sport, mostly on airplanes Tiny map, dry land, 8 other players, domination#11
#11
