What it measures
Open research problems are tracked through computer-checkable constructions and documented mathematical advances. Public first-solver credits are kept distinct from controlled benchmark success rates.
Reading the measurement
A correct final answer, a complete proof and useful research progress are different achievements. Choose the one the task actually grades. Inspect the permitted computation, attempt budget and handling of invalid arguments before comparing models.
Epoch data and evaluation setting
Epoch AI publishes this benchmark in its capability collection. Imported rows retain the published score column and model configuration. Original authors and evaluation providers remain credited in the result references.
Epoch attribution and score coverage
Data is attributed to Epoch AI under CC BY 4.0, with original evaluation providers credited in the linked source snapshots. Scores are imported only where a published mapping identifies the field and its scale; missing or incompatible records are omitted rather than assigned zero.
Published evaluation
Counts of distinct problems with a public first-solver model attribution among Epoch’s five AI-solved records. These are verified public solution records, not scores from a matched full-suite evaluation; attempts, tool budgets and opportunity to solve each problem differ. Retrieval of an already-existing construction can satisfy a problem verifier. The registry contains 49 problems; absent model records are not zero scores. Collaborative human+AI records are a separate measure.
Where this fits in an evaluation suite
Final-answer accuracy, proof credit, and expert-reviewed research progress are distinct signals. A proof-aware evaluation should retain partial work and audit strong claims.
Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.
A reproducible comparison
Counts of distinct problems with a public first-solver model attribution among Epoch’s five AI-solved records. These are verified public solution records, not scores from a matched full-suite evaluation; attempts, tool budgets and opportunity to solve each problem differ. Retrieval of an already-existing construction can satisfy a problem verifier. The registry contains 49 problems; absent model records are not zero scores. Collaborative human+AI records are a separate measure.
Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.
PRIMARY EVIDENCE & FURTHER READING
Reviewed 2026-10-27. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.