datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
NuminaMath-1.5-RL-Verifiable
Dataset Card for NuminaMath-1.5-RL-Verifiable
Dataset Summary
NuminaMath-1.5-RL-Verifiable is a curated subset of the NuminaMath-1.5 dataset, specifically filtered to support reinforcement learning applications requiring verifiable outcomes. This collection consists of 131,063 math word problems from the original dataset that meet strict filtering criteria: all problems have definitive numerical answers, validated problem statements and solutions, and come from… See the full description on the dataset page: https://huggingface.co/datasets/nlile/NuminaMath-1.5-RL-Verifiable.SWE-Bench-Verified-O1-reasoning-high-results
SWE-Bench Verified O1 Dataset
Executive Summary
This repository contains verified reasoning traces from the O1 model evaluating software engineering tasks. Using OpenHands + CodeAct v2.2, we tested O1's bug-fixing capabilities on the SWE-Bench Verified dataset, achieving a 28.8% success rate across 500 test instances.
Overview
This dataset was generated using the CodeAct framework, which aims to improve code generation through enhanced action-based reasoning.… See the full description on the dataset page: https://huggingface.co/datasets/AlexCuadron/SWE-Bench-Verified-O1-reasoning-high-results.Multi-SWE-RL-Verified
Multi-SWE-RL-Verified
Gold-patch-validated subset of
PrimeIntellect/Multi-SWE-RL-Reupload
(ByteDance's Multi-SWE-RL): 2,232 / 4,703 rows across
C, Go, Java, JavaScript, Rust, and TypeScript that produce a clean reward signal end-to-end.
Default dataset of the multiswe_v1 taskset.
Changes vs upstream
Starting from the 4,703-row re-upload:
C++ dropped wholesale — 0/449 rows passed gold-patch validation in pass 1; the images are
broken for scoring, not merely… See the full description on the dataset page: https://huggingface.co/datasets/PrimeIntellect/Multi-SWE-RL-Verified.Big-Math-RL-Verified
Big-Math: A Large-Scale, High-Quality Math Dataset for Reinforcement Learning in Language Models
Big-Math is the largest open-source dataset of high-quality mathematical problems, curated specifically for reinforcement learning (RL) training in language models. With over 250,000 rigorously filtered and verified problems, Big-Math bridges the gap between quality and quantity, establishing a robust foundation for advancing reasoning in LLMs.
Request Early Access to Private… See the full description on the dataset page: https://huggingface.co/datasets/SynthLabsAI/Big-Math-RL-Verified.epoch_ai_swebench_verified
Epoch AI SWE-bench Verified Traces
Complete public trace archives and an analysis-ready Parquet conversion of Epoch AI's SWE-bench Verified evaluations.
Contents
34 published evaluation runs covering 16,456 traces (484 SWE-bench instances per run).
data/: loadable Parquet data, one exact trace per row.
original/: the byte-identical .eval archives published by Epoch AI.
run_metadata/: non-sample files from each .eval archive (header.json, summaries, reductions… See the full description on the dataset page: https://huggingface.co/datasets/naderalfares/epoch_ai_swebench_verified.Scale-SWE-Verified
Scale-SWE-Verified
Gold-patch-validated fork of
AweAI-Team/Scale-SWE
(paper): 17,202 / 20,181 Python issue-resolving tasks
that produce a clean reward signal end-to-end. Default dataset of the scaleswe_v1 taskset.
Changes vs upstream
Validation (ours) removed 2,979 / 20,181 rows (14.8%):
892 rows whose image_url appears in
scale-swe-exclude-images.json.
2,061 rows categorized gold_patch_failure in
scale-swe-validation.jsonl.
15 rows categorized noop_pass… See the full description on the dataset page: https://huggingface.co/datasets/PrimeIntellect/Scale-SWE-Verified.R2E-Gym-Subset-Verified
R2E-Gym-Subset-Verified
Gold-patch-validated subset of
R2E-Gym/R2E-Gym-Subset
(paper). The train split contains
4,522 / 4,578 rows (98.78%) verified scoreable end-to-end: apply the gold patch, run the
upstream /testbed/run_tests.sh baked into the row's image, check the parsed outcomes against
expected_output_json.
Changes vs upstream
Validation-only subset — our passes, run in fresh sandboxes per row: one full pass at
concurrency 200, then a 10× retry pass over… See the full description on the dataset page: https://huggingface.co/datasets/PrimeIntellect/R2E-Gym-Subset-Verified.SWE-rebench-V2-Filtered-Verified
SWE-rebench-V2-Filtered-Verified
Filtered and gold-patch-verified subset of Nebius's
SWE-rebench-V2
(paper): 6,272 / 32,079 freshly-mined GitHub PR tasks
across 17 languages. Default dataset of the swerebench_v2_v1 taskset.
Changes vs upstream
Filtered (selection — the bulk of the cut):
Upstream's own per-row LLM-judge metadata: difficulty labeled easy/medium/hard, judge grade
code == "A" (clearly solvable), intent_completeness == "complete", no detected_issues… See the full description on the dataset page: https://huggingface.co/datasets/PrimeIntellect/SWE-rebench-V2-Filtered-Verified.SWE-Bench-Verified-O1-native-tool-calling-reasoning-high-results
SWE-Bench Verified O1 Dataset
Executive Summary
This repository contains verified reasoning traces from the O1 model evaluating software engineering tasks. Using OpenHands + CodeAct v2.2, we tested O1's bug-fixing capabilities using their native tool calling capabilities on the SWE-Bench Verified dataset, achieving a 45.8% success rate across 500 test instances.
Overview
This dataset was generated using the CodeAct framework, which aims to improve code… See the full description on the dataset page: https://huggingface.co/datasets/AlexCuadron/SWE-Bench-Verified-O1-native-tool-calling-reasoning-high-results.SciCode-Verified
SciCode-Verified
SciCode-Verified is the corrected, human-verified release of the
SciCode scientific-code-generation benchmark.
A problem-by-problem audit identified 263 defects in the 65-problem SciCode test split and
corrected every confirmable defect. The released evaluation set contains 64 main problems and
287 scored subproblems; one original problem is excluded because its specification does not
determine a unique, verifiable answer.
Paper: SciCode-Verified: How Benchmark… See the full description on the dataset page: https://huggingface.co/datasets/shhu2001/SciCode-Verified.SWE-Lego-Real-Data-Verified
SWE-Lego-Real-Data-Verified
Gold-patch-validated subset of
PrimeIntellect/SWE-Lego-Real-Data
(itself a fixed fork of SWE-Lego's real-data split). The
resolved split contains 4,323 / 4,432 rows (97.54%) verified scoreable end-to-end: apply
test_patch, apply the gold patch, run the row's test_cmd in its image, require every
F2P/P2P test to report PASSED.
Changes vs upstream
Validation-only subset — our passes: one full pass at concurrency 200, then a 10× retry… See the full description on the dataset page: https://huggingface.co/datasets/PrimeIntellect/SWE-Lego-Real-Data-Verified.TACO-verified
Introduction
This dataset contains verified solutions from the TACO dataset's training set. Solutions that fail to pass all the test cases are removed. Problems with no correct solution are also removed.
The solutions were executed on Intel E5-2620 v3 CPUs with the execution timeout set to 10 seconds.
Statistics in the training set
Dataset
# Problems
# Solutions
TACO
25443
1468722
TACO-verified
12898
1043251
Correct Ratio
50.69 %
71.03 %… See the full description on the dataset page: https://huggingface.co/datasets/likaixin/TACO-verified.verifiable-code-reasoning
Verifiable Code Reasoning
Execution-verified Python problems with chain-of-thought
Sandbox-checked solutions · Multi-test unit checks · Deduplicated instances · Training-ready sft_text
Overview
Verifiable Code Reasoning is a large-scale dataset of Python coding problems where every kept solution has passed sandboxed unit tests.
Unlike scraped contest dumps or unverified LLM traces, an example enters this release only if:
a reference… See the full description on the dataset page: https://huggingface.co/datasets/smshahbaj/verifiable-code-reasoning.terminal-bench-2-verified
Terminal-Bench 2.0 Verified: Instruction & Environment Fix Version
中文版本
We conducted a comprehensive review of the entire Terminal-Bench 2.0 dataset and identified various issues. Both GLM-5 and Step 3.5-Flash were evaluated using this verified version.
This modified version addresses environment and instruction issues we discovered in Terminal-Bench 2.0. It includes two types of fixes:
Environment Fixes: Updated Dockerfiles and instructions to support Claude Code Agent runtime… See the full description on the dataset page: https://huggingface.co/datasets/harithoppil/terminal-bench-2-verified.VeriLoop-E2-Evaluation-Evidence
VeriLoop E2 Evaluation Evidence
Public evaluation evidence for VeriLoop E2 across nine code, agentic, mathematical, and scientific reasoning benchmarks.
This dataset repository is the canonical public evidence layer for the reported benchmark results of VeriLoop E2, a post-trained model based on Qwen 3.8-27B. It is designed to separate headline benchmark reporting from the underlying auditable artifacts required to inspect, reproduce, and verify those results.
The repository… See the full description on the dataset page: https://huggingface.co/datasets/tsinghua-sigs-robot-lab/VeriLoop-E2-Evaluation-Evidence.medical-o1-verifiable-problem
Introduction
This dataset features open-ended medical problems designed to improve LLMs' medical reasoning. Each entry includes a open-ended question and a ground-truth answer based on challenging medical exams. The verifiable answers enable checking LLM outputs, refining their reasoning processes.
For details, see our paper and GitHub repository.
Citation
If you find our data useful, please consider citing our work!
@misc{chen2024huatuogpto1medicalcomplexreasoning… See the full description on the dataset page: https://huggingface.co/datasets/FreedomIntelligence/medical-o1-verifiable-problem.lean-verifier-formalizations
Lean Verifier Formalizations
A dataset of Lean 4 theorem-proving tasks for evaluating agentic coding harnesses. Each row pairs a formal task_statement (with the reference proof body removed) against a real Lean 4 repository, plus the informal_excerpt/informal_source_text describing what the theorem claims, permitted_axioms for the verifier, and provenance fields (repo_url, repo_commit_sha, license) tracing back to the source project.
Sources
Every row is pulled… See the full description on the dataset page: https://huggingface.co/datasets/Pradheep1647/lean-verifier-formalizations.verified-defi-datasets
Verified Solana Sealevel & Anchor Program Optimization Fine-Tuning Corpus
Dataset Description
High-density, verified AI fine-tuning dataset in ALPACA format.
Domain: Solana Sealevel & Anchor Program Optimization
Verified Records: 3
Estimated Tokens: 339
Quality QA Score: 99.0%
Monetization Status: Direct Zero-Gas Web3 & HuggingFace Distribution
SWE-rebench-V2-Filtered-Easy-Verified
SWE-rebench-V2-Filtered-Easy-Verified
Easy slice of
PrimeIntellect/SWE-rebench-V2-Filtered-Verified:
rows whose upstream LLM-judge difficulty is easy (implementation-time estimate < 15 min).
Useful as a lower-variance starting pool for RL curricula.
Changes vs upstream
Pure slice of the Filtered-Verified set — it inherits every filter and verification pass from
the parent (see its card), including the pass-2 flaky removal, no-edit pass, and repo/image
blocklists… See the full description on the dataset page: https://huggingface.co/datasets/PrimeIntellect/SWE-rebench-V2-Filtered-Easy-Verified.simpleqa-verified-multilingual
SimpleQA Verified Multilingual
Multilingual translations of SimpleQA Verified, a 1,000-prompt factuality benchmark from Google DeepMind that evaluates short-form parametric knowledge (facts stored in model weights).
Source: google/simpleqa-verified (eval split, 1,000 examples)
Languages
Config
Language
Examples
ces
Czech
100
dan
Danish
100
deu
German
1,000
fra
French
100
ita
Italian
100
nld
Dutch
100
pol
Polish
100
spa
Spanish
100
More to… See the full description on the dataset page: https://huggingface.co/datasets/ellamind/simpleqa-verified-multilingual.nemotron-nano-30b-miniswe-swebench-verified
Nemotron Nano 30B + mini-swe-agent SWE-bench Verified Trajectories
Agent trajectories from running NVIDIA Nemotron 3 Nano 30B A3B (MoE, 8B active params) on SWE-bench Verified using mini-swe-agent.
⚠️ Incomplete Run
This benchmark was terminated early due to poor performance. The model struggled with the agentic coding task.
Model Information
Attribute
Value
Model
NVIDIA Nemotron 3 Nano 30B A3B
Architecture
MoE (30B total, 8B active)
Serving
vLLM… See the full description on the dataset page: https://huggingface.co/datasets/pankajmathur/nemotron-nano-30b-miniswe-swebench-verified.afd_mix_p50_verified
Aha-Flow Distillation: Flow Markers Matter in LLM Reasoning
Overview
We identify the Flow Moment, a new confident reasoning format in LLM reasoning, and we explore the influence of this format on On-Policy Self-Distillation (OPSD). We propose Aha-Flow Distillation, a dual-mode extension of OPSD that combines Aha branch and Flow branch to distill the student model. This training on Qwen3-8B model only takes ~30 minutes on 8×H20 and peaks within 100… See the full description on the dataset page: https://huggingface.co/datasets/Xiaodong/afd_mix_p50_verified.Lego-RL-SWE-Bench-Verified
Lego-RL-SWE-Bench-Verified
The 500 SWE-bench Verified instances as ready-to-run harbor RL environments —
the exact evaluation set behind every SWE-bench Verified number in
LEGO-RL, packaged the same way as the training
set Lego-X/Lego-RL-2699 so
one trainer reads both.
Two parallel views of the same 500 instances:
View
Path
What it is
Official SWE-bench records
swebench_verified_official_500/
The upstream princeton-nlp/SWE-bench_Verified rows, verbatim
Harbor RL… See the full description on the dataset page: https://huggingface.co/datasets/Lego-X/Lego-RL-SWE-Bench-Verified.afd_mix_p100_verified
Aha-Flow Distillation: Flow Markers Matter in LLM Reasoning
Overview
We identify the Flow Moment, a new confident reasoning format in LLM reasoning, and we explore the influence of this format on On-Policy Self-Distillation (OPSD). We propose Aha-Flow Distillation, a dual-mode extension of OPSD that combines Aha branch and Flow branch to distill the student model. This training on Qwen3-8B model only takes ~30 minutes on 8×H20 and peaks within 100… See the full description on the dataset page: https://huggingface.co/datasets/Xiaodong/afd_mix_p100_verified.MAPS_Verified
Dataset Card for Multilingual Benchmark for Global Agent Performance and Security
This is the first Multilingual Agentic AI Benchmark for evaluating agentic AI systems across different languages and diverse tasks. Benchmark enables systematic analysis of how agents perform under multilingual conditions. This dataset contains 550 instances for GAIA, 660 instances for ASB, 737 instances for Maths, and 1100 instances for SWE. Each task was translated into 10 target languages resulting… See the full description on the dataset page: https://huggingface.co/datasets/Fujitsu-FRE/MAPS_Verified.Doc2Feat-bench_Verified
Dataset Summary
NoCode-bench Verified is subset of NoCode-bench, a dataset that tests systems’ no-code feature addition ability automatically.
Languages
The text of the dataset is primarily English, but we make no effort to filter or otherwise clean based on language type.
Dataset Structure
An example of a SWE-bench datum is as follows:
repo: (str) - The repository owner/name identifier from GitHub.
instance_id: (str) - A formatted instance… See the full description on the dataset page: https://huggingface.co/datasets/Doc2Feat-bench/Doc2Feat-bench_Verified.verified_wiki_historian_the_beatles_anthology_dataset_active
Verified-Wiki-Historian: The Beatles Anthology
Verified-Wiki-Historian (The Beatles Anthology) is a refined, citation-grounded instruction dataset for Beatles-specific historical question answering, summarization, and supervised fine-tuning.
This dataset is a cleaned and rebuilt refinement of:
Mungus451/verified_wiki_historian_the_beatles_anthology_dataset_active
The current release contains 4,000 instruction records focused on Beatles history, recording sessions, release… See the full description on the dataset page: https://huggingface.co/datasets/Mungus451/verified_wiki_historian_the_beatles_anthology_dataset_active.VeriTime
VeriTime: Time Series Reasoning via Process-Verifiable Thinking Data Synthesis and Scheduling for Tailored LLM Reasoning
This is the dataset associated with our paper:
Time Series Reasoning via Process-Verifiable Thinking Data Synthesis and Scheduling for Tailored LLM Reasoning
Jiahui Zhou, Dan Li, Boxin Li, Xiao Zhang, Erli Meng, Lin Li, Zhuomin Chen, Jian Lou, See-Kiong Ng
ICML 2026 | Paper
Dataset Construction Pipeline: TSRgen
TSRgen is an… See the full description on the dataset page: https://huggingface.co/datasets/HayleyZhou1113/VeriTime.verifiable-corpus
verifiable-corpus
This is the corpus from "Learning on the Job: Test-Time Curricula for Targeted Reinforcement Learning".
Code: https://github.com/jonhue/ttc
Introduction
We study how large language models (LLMs) can continually improve at reasoning on their target tasks at test-time. We propose an agent that assembles a task-specific curriculum, called test-time curriculum (TTC-RL), and applies reinforcement learning to continue training the model for its target task.… See the full description on the dataset page: https://huggingface.co/datasets/lasgroup/verifiable-corpus.execution-verified-codework
Execution-Verified CodeWork
Only code that passes the tests ships
Sandbox-executed · ≥6 unit tests · implement / repair / harden · instance-deduplicated
One-sentence pitch
Training traces for writing, fixing, and hardening Python functions — every kept solution was actually run against unit tests and passed.
What you get
Field
Role
kind
implement · repair · harden
problem
Clear developer task
reasoning
Numbered… See the full description on the dataset page: https://huggingface.co/datasets/smshahbaj/execution-verified-codework.