Reasoning Core ◉

Procedural reasoning data for language-model pre-training, post-training, evaluation, and RL.

Reasoning Core generates verifiable textual tasks across first-order logic, formal mathematics with Lean and TPTP, planning, algorithms, syntax, and more. Use it as a Python library, generate datasets at scale, or plug it into modern reinforcement-learning environments.

More than 10B tokens of pre-generated data are available in the 🤗 Reasoning Core dataset collection.

Tasks target compact, canonical answers and expose task-native scorers—clean targets for supervised fine-tuning (SFT), with deterministic rewards for RL and evaluation. Task development is principled, data-driven, and agent-friendly: task diagnostics provide reproducible audits of transfer, solvability, and shortcut resistance so every task can earn its place.

Quickstart

uv pip install reasoning-core

from reasoning_core import get_task, score_answer

task = get_task("arithmetics")
example = task.generate_example()

print(example.prompt)
assert score_answer(example.answer, example) == 1

Representative example

Reasoning Core includes compilation-checked formal reasoning tasks such as lean_candidate_compilation:

Prompt

Does this Lean 4 tactic body close the theorem?
The answer is True or False.

THEOREM:
theorem ex (p2 p4 : Prop) : p2 → (p2 ∨ p4) := by
  ?

CANDIDATE:
linarith

Answer: False

Browse all 50 task examples.

Task catalogue

GALLERY (names link to gallery examples)

arithmetics · math_word_problem · equation_system · lean_missing_line · lean_candidate_compilation · planar_geometry_relations · metamath_entailment · metamath_core_select · lambda_reduction · rewrite_system · unification_entailment · most_probable_evidence · most_probable_outcome · logic_nli · logic_formalization · multistep_nli · defeasible_nli · multistep_evidence_retrieval · multistep_abduction · logic_qa · planning · set_missing_element · set_expression · sequential_induction · qualitative_reasoning · grid_navigation · reference_tracking · belief_tracking · coreference · constraint_satisfaction · graph_pathfinding · graph_successors · graph_dependencies · regex_following · regex_reasoning · analogical_case_matching · parsing_derivation · syntax_error_detection · constrained_continuation · table_qa · table_equivalence · table_statistics · string_transduction · game_best_move · game_forced_win · qualitative_causal_reasoning · code_analysis · code_runnability · code_execution · program_synthesis

Task authoring guidelines

A task authoring guide describes the interface and guidelines.
The task diagnostics describes fast bust robust task influence measurement for task validation.
TASK_AUTHORING_GUIDE TASK_DIAGNOSTICS

Ecosystem and integrations

• Prime Intellect — install Reasoning Core from the Environments Hub for evaluation and RL workflows.
• OpenReward — run Reasoning Core as an OpenReward-compatible environment.
• OpenEnv — explore the interactive Reasoning Core OpenEnv on Hugging Face Spaces.
• reasoning-gym — mix Reasoning Core and reasoning-gym tasks through either library's interface.
• SynLogic — generate SynLogic games through the same task API, backed by SynLogic's native verifiers.

See the integration guide for runnable examples.

Generate datasets at scale

Go from a single example to large pre-training, post-training, and evaluation corpora with balanced difficulty, token budgets, and verifiable answers. The generation pipeline supports parallel workers, resumable jobs, and JSONL shards and scalable postprocessing scripts ready for Hugging Face Datasets.

Citation and paper

@article{reasoningcore2026,
  title={Reasoning Core: A Scalable Procedural Data Generation Suite for Symbolic Pre-training and Post-Training},
  author={Lacombe, Valentin and Quesnel, Valentin and Sileo, Damien},
  journal={arXiv preprint arXiv:2603.02208},
  year={2026},
  url={https://arxiv.org/abs/2603.02208}
}

https://arxiv.org/abs/2603.02208
Contact: damien.sileo@inria.fr