B 1.1

Base language model benchmark

BananaMindBench Leaderboard

Text-completion performance on 350 BananaMind Base Bench 1.1 examples, ranked by fixed-item Overall Elo.

Maximum parameter size for submissions: 250M.

Dataset Official runner Submit a model

350 examples 7 categories 4 continuations per example Mean conditional log-likelihood Higher Elo is better

Search Organization

Model size

Official results

Base Bench 1.1

Overall Elo is the benchmark's weighted fixed-item logistic rating. It is not an average of the category Elo values.

Results use plain text and continuation likelihood. No chat template, role labels, sampling, or free-form generation is used.