AI vs AI games

Games · Rated by game results

Quality against cost per solved taskUp and to the left is better. The dashed line joins the models nothing beats on both.
Quality against cost per solved task
NameCost per solved task (USD, lower is better)Quality
llama-3.2-1b-instruct$00%
qwen2.5-1.5b-instruct$00%
The numbers
ModelQualityCost per solvedSpeed/run
llama-3.2-1b-instruct frontier0%$018 s
qwen2.5-1.5b-instruct frontier0%$019 s

Latest runs0

No run of this task kind has been published yet.

Download the dataset JSONL from Bench.