GooseReason-0.7M is a large-scale Reinforcement Learning with Verifiable Rewards (RLVR) dataset released by NVIDIA in 2026, associated with the paper "Golden Goose: A Simple Trick to Synthesize Unlimited RLVR Tasks from Unverifiable Internet Text", aiming to address the bottleneck of scarce verifiable training data by synthesizing unlimited RLVR tasks from rich reasoning-oriented internet text via the Golden Goose pipeline.

The dataset contains over 700,000 tasks covering mathematics, programming, and general science domains, with a data scale at the 0.7M level. It is primarily used to train the GooseReason-4B-Instruct model, which achieves new state-of-the-art results for 4B-parameter models across 15 benchmarks encompassing mathematics, programming, STEM reasoning, instruction following, and logic puzzles. This dataset is intended solely for research and development purposes.

Dataset Composition

• Math Domain: Contains 235,836 samples sourced from AoPS-Instruct (approximately 600,000 QA pairs from the Art of Problem Solving forum, primarily consisting of olympiad-level mathematics problems).
• Code Domain: Contains 281,793 samples sourced from rStar-Coder (approximately 418,000 programming problems from competitive programming platforms, using the synthetic_sft split, containing only questions and teacher model solutions, without test cases).
• STEM Domain: Contains 155,496 samples sourced from MegaScience (approximately 650,000 QA pairs from about 12,000 university-level science textbooks, covering physics, biology, chemistry, medicine, computer science, mathematics, and economics).

Citation

@article{lu2026goldengoose,
  title={Golden Goose: A Simple Trick to Synthesize Unlimited RLVR Tasks from Unverifiable Internet Text},
  author={Lu, Ximing and Acuna, David and Jung, Jaehun and Hu, Jian and Zhang, Di and Diao, Shizhe and Zou, Yunheng and Zhang, Shaokun and Cui, Brandon and Liu, Mingjie and Kim, Hyunwoo and Ammanabrolu, Prithviraj and Kautz, Jan and Dong, Yi and Choi, Yejin},
  journal={arXiv preprint arXiv:2601.22975},
  year={2026}
}