Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Lemma Judging research benchmarks · SOTA2 Research

Lemma Judging

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

DeepTheorem sampled perturbation (test)

Llama-3.1 (RULES)

90.9Exact Match Accuracy

12

Feb 26, 2026

IMO-Lemma (IL) (test)

Llama-3.1 (RULES)

93.6Exact Match Accuracy

12

Feb 26, 2026

Llama-3.1 (GRPO)

86.5Exact Match Accuracy

12

Feb 26, 2026

NaturalProofs (test)

Llama-3.1 (RULES)

93.2Exact Match Accuracy

12

Feb 26, 2026

Popular datasets

DeepTheorem NaturalProofs NLPS IMO-Lemma

Follow for updates

Join SOTA2 Contact