What it measures

A mathematical-reasoning benchmark built from Putnam competition problems, with original questions and functional variations that test whether a model can adapt its solution when a problem changes. Problem accuracy is separate from teacher-forced token prediction.

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.

Published evaluation

Percent of 522 original problems solved. Teacher-forced token accuracy (TFA) is a different metric and is not presented as problem-solving success.

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

Percent of 522 original problems solved. Teacher-forced token accuracy (TFA) is a different metric and is not presented as problem-solving success.

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

  1. Putnam-AXIOM: A Functional and Static Benchmark for Measuring Higher Level Mathematical Reasoning in LLMs — published result table ↗

    Gulati, Aryan, Miranda, Brando, Chen, Eric, Xia, Emily, Fronsdal, Kai, Dumont, Bruno et al. · 2025

    All 8 authors

    Gulati, Aryan, Miranda, Brando, Chen, Eric, Xia, Emily, Fronsdal, Kai, Dumont, Bruno, Obbad, Elyas, Koyejo, Sanmi

Reviewed 2026-10-26. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.