MLE-bench: Results as Published

MLE-bench tests whether an AI agent can do machine learning engineering end to end. Its tasks are 75 Kaggle competitions: the agent must prepare the data, train models, run experiments and submit predictions. Each submission is placed against the human leaderboard of that competition, and the headline score is the share of competitions in which the agent reaches at least a bronze medal.

Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.