MLE-bench

Kaggle-style machine learning engineering tasks covering training, data preparation, and experiments.

1rows
bronze_medal_rateprimary metric
2024-10-10sampled

Metadata

ID
openai_mle_bench
Category
Self Improvement
Release
2024-10-10

Metrics

Bronze Medal Rate

Latest Results

OpenAI-reported best-performing setup from the original benchmark release. The score is the share of competitions where the setup achieved at least Kaggle bronze level.

Rank Subject Bronze Medal Rate Model Match Provenance Sampled
1 o1-preview + AIDE scaffolding 16.9% o1-preview
openai-o1-preview
 		Imported 2024-10-10