Note RL dataset for olympiads Math. Contains hints that are used for difficult samples. Details in https://arxiv.org/pdf/2507.10628.

Note answers were provided by R1 then extracted.

Note needs a judge

Note A collection of 235K samples for math curr learning.

Note A large-scale dataset consisting of approximately 100,000 algebraic word problems. MCQ with single correct answer. Can be verified via string matching with the final answer.

Note contains math and code

Note Math split contains 105k samples

Note 34K samples of math

Note sample some from the medium and hard sets. (hard is 1% which is already 2.3M samples...)

Note 270k math problems

Note high quality, verfiable, challenging and diverse math dataset for training math reasoning model using reinforcement leraning. This dataset contains 49K math problems and answer sourced from NuminaMath and DeepScaler-Preview applying filtering rules to exclude unsuitable data (e.g., multiple sub-questions, multiple-choice, true/false, long and complex answers, proof, figure) this dataset was used to train AceReason-Nemotron models, which achieve strong results on math benchmark such as AIME24 an