MathArena · ArXivMath
MathArena / ETH Zurich / Reasoning · New problems drawn from recent mathematics papers
Across e/accUnder observation
Evidence budgetNot scored
Upstream data timeUnverified
Last successful syncCollection not started
What it measures and how
Read raw results for each task set and version separately. Do not use the aggregate Expected performance score fitted across task sets.
How this evidence is used
Under observation: terms for displaying leaderboard results remain unclear. Results are not collected for ranking yet.
Limitations and attribution
Under observation: terms for displaying leaderboard results remain unclear. Results are not collected for ranking yet.
Data license: Code is MIT. Data includes CC BY-NC-SA 4.0 restrictions; distinguish task datasets from leaderboard results.
Results are published by MathArena / ETH Zurich. Raw scores and the e/acc consensus index use different scales and cannot be added directly.