GooseReason-0.7M is a large-scale RLVR dataset with over 0.7 million tasks across mathematics, programming, and general scientific domains, synthesized by the Golden Goose pipeline. It is used to train GooseReason-4B-Instruct, which achieves new state-of-the-art results among 4B-Instruct models across 15 diverse benchmarks, spanning mathematics, programming, STEM reasoning, instruction following, and logical puzzles.

This dataset is for research and development only.

Golden Goose

Scaling up RLVR is bottlenecked by the scarcity of verifiable training data, where improvements increasingly saturate after prolonged training on existing datasets. Golden Goose is a simple, scalable pipeline that synthesizes unlimited RLVR tasks from reasoning-rich but unverifiable internet text—corpora such as science textbooks, Olympiad math forums, and cybersecurity web scrapes that were previously excluded from RLVR data construction due to the difficulty of automatic verification.

The key idea: given a source text S, we prompt an LLM to identify a contiguous span t of crucial reasoning steps and replace it with a [MASK] token, constructing a masked context S_mask. Treating t as the ground-truth answer, the LLM then generates a set of diverse, plausible distractors D = {d₁, ..., dₖ} that are similar in style and length to the removed span yet incorrect in context, forming a multiple-choice question: Q = (S_mask, {t} ∪ D)

Verification during RL simply checks whether the model's prediction matches the ground-truth option—no external judge or test execution needed. This formulation unlocks reasoning-rich corpora that were previously unusable for RLVR: Olympiad-level theorem proving from AoPS-Instruct, free-form textbook QA from MegaScience, and coding problems without test cases from rStar-Coder.

GooseReason-0.7M Dataset

Using the Golden Goose pipeline, we synthesize GooseReason-0.7M, a large-scale RLVR dataset with over 0.7 million tasks spanning mathematics, programming, and general scientific domains. The dataset is constructed from the following source corpora:

The data mixing ratio used to train GooseReason-4B-Instruct is 55% ProRL data, 15% GooseReason-0.7M Math, 15% GooseReason-0.7M Code, and 15% GooseReason-0.7M STEM.

Data Format

Each record contains three fields:

Example record:

{
  "question": "You are given a math problem and its solution, with some steps replaced by [MASK]...\n\n**Question:**\n...\n\n**Solution:**\n...\n[MASK]\n...",
  "options": [
    "6. Because QR is a radical axis, the point E has equal power...",
    "6. Using power of a point, E has equal power to the circle...",
    "..."
  ],
  "answer": "E"
}

Citation

If you find this dataset or the Golden Goose paper helpful, please cite:

@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}
}