datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
OpenCodeInstruct
OpenCodeInstruct: A Large-scale Instruction Tuning Dataset for Code LLMs
Dataset Description
We introduce OpenCodeInstruct, the largest open-access instruction tuning dataset, comprising 5 million diverse samples. OpenCodeInstruct is designed for supervised fine-tuning (SFT).
Technical Report - Discover the methodology and technical details behind OpenCodeInstruct.
Github Repo - Access the complete pipeline used to perform SFT.
This dataset is ready for… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/OpenCodeInstruct.OpenCodeReasoning
OpenCodeReasoning: Advancing Data Distillation for Competitive Coding
Data Overview
OpenCodeReasoning is the largest reasoning-based synthetic dataset to date for coding, comprises 735,255 samples in Python across 28,319 unique competitive programming
questions. OpenCodeReasoning is designed for supervised fine-tuning (SFT).
Technical Report - Discover the methodology and technical details behind OpenCodeReasoning.
Github Repo - Access the complete pipeline used to… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/OpenCodeReasoning.nvidia-OpenCodeInstruct-refined
nvidia-OpenCodeInstruct-refined
A strictly quality-filtered subset of nvidia/OpenCodeInstruct (5M examples). This is a strict subset of EER6/nvidia-OpenCodeInstruct-broad.
Filtering criteria
Both conditions must be satisfied:
Criterion
Threshold
LLM judge min score
= 5 (out of 5)
Unit test pass rate (average_test_score)
= 1.0
LLM judge min score is the minimum across all three dimensions in the llm_judgement field:
requirement_conformance — does the… See the full description on the dataset page: https://huggingface.co/datasets/EER6/nvidia-OpenCodeInstruct-refined.OpenCodeReasoning-2
OpenCodeReasoning-2: A Large-scale Dataset for Reasoning in Code Generation and Critique
Dataset Description
OpenCodeReasoning-2 is the largest reasoning-based synthetic dataset to date for coding, comprising 1.4M samples in Python and 1.1M samples in C++ across 34,799 unique competitive programming questions.
OpenCodeReasoning-2 is designed for supervised fine-tuning (SFT) tasks of code completion and code critique.
Github Repo - Access the complete pipeline used to… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/OpenCodeReasoning-2.Nemotron-SFT-OpenCode-v1
Dataset Description:
Nemotron-SFT-OpenCode-v1 is an agentic instruction tuning dataset that enhances the ability of Large Language Models (LLMs) to operate within the OpenCode Command Line Interface (CLI) framework and instills simple capabilities such as tool calling and agent skills.
This dataset is ready for commercial/non-commercial use.
Dataset Subsets:
Nemotron-SFT-OpenCode-v1 contains the following subsets, where the questions and agent skills are synthetically… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Nemotron-SFT-OpenCode-v1.OpenCodeReasoning_messages
This is a transformation of the nvidia/OpenCodeReasoning dataset into a format that is more easily digestible by trainers.
OpenCodeReasoning: Advancing Data Distillation for Competitive Coding
Data Overview
OpenCodeReasoning is the largest reasoning-based synthetic dataset to date for coding, comprises 735,255 samples in Python across 28,319 unique competitive programming
questions. OpenCodeReasoning is designed for supervised fine-tuning (SFT).
Technical Report -… See the full description on the dataset page: https://huggingface.co/datasets/sealad886/OpenCodeReasoning_messages.OpenCodeReasoning-2
OpenCodeReasoning-2: A Large-scale Dataset for Reasoning in Code Generation and Critique
Dataset Description
OpenCodeReasoning-2 is the largest reasoning-based synthetic dataset to date for coding, comprising 1.4M samples in Python and 1.1M samples in C++ across 34,799 unique competitive programming questions.
OpenCodeReasoning-2 is designed for supervised fine-tuning (SFT) tasks of code completion and code critique.
Github Repo - Access the complete pipeline used to… See the full description on the dataset page: https://huggingface.co/datasets/cublya/OpenCodeReasoning-2.OpenCodeInstruct
OpenCodeInstruct: A Large-scale Instruction Tuning Dataset for Code LLMs
Dataset Description
We introduce OpenCodeInstruct, the largest open-access instruction tuning dataset, comprising 5 million diverse samples. OpenCodeInstruct is designed for supervised fine-tuning (SFT).
Technical Report - Discover the methodology and technical details behind OpenCodeInstruct.
Github Repo - Access the complete pipeline used to perform SFT.
This dataset is ready for… See the full description on the dataset page: https://huggingface.co/datasets/Jeremydh911/OpenCodeInstruct.NMT-opencode
OpenCodeReasoning: Advancing Data Distillation for Competitive Coding
Data Overview
OpenCodeReasoning is the largest reasoning-based synthetic dataset to date for coding, comprises 735,255 samples in Python across 28,319 unique competitive programming
questions. OpenCodeReasoning is designed for supervised fine-tuning (SFT).
Technical Report - Discover the methodology and technical details behind OpenCodeReasoning.
Github Repo - Access the complete pipeline used to… See the full description on the dataset page: https://huggingface.co/datasets/Compumacy/NMT-opencode.opencode_reasoning_filtered
🧠 OpenCode Reasoning (Filtered)
Author: Parvesh Rawal — XenArcAILicense: Inherits from NVIDIA OpenCodeReasoningVersion: Filtered & Structured VariantTotal Examples: 567,850Total Size: 9GB (compressed)
🔍 Overview
This dataset is a curated and cleaned version of split_0 from nvidia/OpenCodeReasoning, optimized for code-level reasoning tasks and instruction tuning.
It’s designed to enhance logic understanding and multistep problem solving for LLMs.
📁 Features… See the full description on the dataset page: https://huggingface.co/datasets/Parveshiiii/opencode_reasoning_filtered.Qwen3-Coder-Next-OpenCode-Preference
Dataset Card — OpenCode Rejection Sampling (Preference)
Overview
This dataset contains 10,920 preference pairs for preference-based training (DPO, KTO, SimPO, ORPO, etc.) on competitive programming tasks. Each pair consists of:
Chosen: a candidate solution that passes 100% of test cases
Rejected: a candidate solution that fails, with a fine-grained rejection type label
Pairs are produced via rejection sampling with Qwen3-Coder-Next: 8 candidate solutions are… See the full description on the dataset page: https://huggingface.co/datasets/zake7749/Qwen3-Coder-Next-OpenCode-Preference.HyperThink-X-Nvidia-Opencode-Reasoning-200K
🔮 HyperThink
HyperThink is a premium, best-in-class dataset series capturing deep reasoning interactions between users and an advanced Reasoning AI system. Designed for training and evaluating next-gen language models on complex multi-step tasks, the dataset spans a wide range of prompts and guided thinking outputs.
🚀 Dataset Tiers
HyperThink is available in three expertly curated versions, allowing flexible scaling based on compute resources and training goals:… See the full description on the dataset page: https://huggingface.co/datasets/Sashvat/HyperThink-X-Nvidia-Opencode-Reasoning-200K.nvidia-OpenCodeInstruct-broad
nvidia-OpenCodeInstruct-broad
A quality-filtered subset of nvidia/OpenCodeInstruct (5M examples).
Filtering criteria
Both conditions must be satisfied:
Criterion
Threshold
LLM judge min score
>= 4 (out of 5)
Unit test pass rate (average_test_score)
>= 0.8
LLM judge min score is the minimum across all three dimensions in the llm_judgement field:
requirement_conformance — does the code do what the instruction asked?
logical_correctness — is the… See the full description on the dataset page: https://huggingface.co/datasets/EER6/nvidia-OpenCodeInstruct-broad.opencode-agentic-mini
OpenCode Agentic Dataset
A high-quality training dataset of real-world agentic coding sessions extracted from OpenCode CLI usage. Contains multi-turn conversations with tool calls, reasoning traces, and code editing across 45+ diverse software projects.
Dataset Summary
Split
Examples
Sessions
Train
18,549
~456
Validation
1,001
~51
Total
19,550
507
Estimated tokens: ~153M
Average tokens per example: ~7,800
Data Format
Each… See the full description on the dataset page: https://huggingface.co/datasets/Petrouil/opencode-agentic-mini.nvidia_opencodeinstruct-score-ge-0.9
dataset original: https://huggingface.co/datasets/nvidia/OpenCodeInstruct
filtered average_test_score >= 0.9
Dataset({
features: ['id', 'input', 'output', 'domain', 'generation_algorithm', 'llm_judgement', 'unit_tests', 'tests_execution_status', 'average_test_score'],
num_rows: 2194336
})
OpenCodeReasoning-Cleaned
OpenCodeReasoning-Cleaned
OpenCodeReasoning-Cleaned is a filtered and reformatted subset derived from nvidia/OpenCodeReasoning, prepared for streamlined supervised fine-tuning (SFT) of code-reasoning language models.
Dataset Summary
This dataset takes the first 100,000 rows of nvidia/OpenCodeReasoning, strips it down to only the two columns needed for instruction/response-style fine-tuning, renames them for clarity, cleans the data by removing null values and… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/OpenCodeReasoning-Cleaned.nemotron-sft-opencode-filtered-250k-v0.1Base dataset: nvidia/Nemotron-SFT-OpenCode-v1
Exact-deduplication + Filtered for the following categories:
code_explanation
code_generation
planning_and_task_structuring
reasoning
code_debugging
advice
brainstorming
code_review
unit_test_generation
summarization
code_refactoring
rewriting_and_editing
OpenCodeReasoning-Cleaned-GRPO
OpenCodeReasoning-Cleaned-GRPO
Deep-cleaned for GRPO/RL training | 500 examples | 525 bugs fixed
📋 Dataset Description
Code reasoning and critique prompts extracted from NVIDIA's OpenCodeReasoning-2 dataset. This cleaned version removes formatting artifacts, HTML tags, and whitespace issues that would degrade GRPO/RL training quality.
Original source: nvidia/OpenCodeReasoning-2 by NVIDIA
📊 Cleaning Statistics
Metric
Value
Original… See the full description on the dataset page: https://huggingface.co/datasets/Eyght/OpenCodeReasoning-Cleaned-GRPO.Alizee-OpenCodeReasoning-Phase3-1.4M
🚀 Alizee OpenCodeReasoning Phase 3 Conformant Dataset - 1.2M Examples
📊 Dataset Summary
This is a fully conformant version of the Phase 3 dataset, processed to strictly follow the specification with clean separation between data and formatting tags. Contains 1.2 million high-quality Python code examples with synthetic prompts and concise reasoning chains.
Key Improvements
✅ 100% Conformant to Phase 3 specification
✅ Synthetic prompts generated from code… See the full description on the dataset page: https://huggingface.co/datasets/DUKEAI/Alizee-OpenCodeReasoning-Phase3-1.4M.Nemotron-SFT-OpenCode-v1
Dataset Description:
Nemotron-SFT-OpenCode-v1 is an agentic instruction tuning dataset that enhances the ability of Large Language Models (LLMs) to operate within the OpenCode Command Line Interface (CLI) framework and instills simple capabilities such as tool calling and agent skills.
This dataset is ready for commercial/non-commercial use.
Dataset Subsets:
Nemotron-SFT-OpenCode-v1 contains the following subsets, where the questions and agent skills are… See the full description on the dataset page: https://huggingface.co/datasets/nick007x/Nemotron-SFT-OpenCode-v1.grpo-opencoder-50k
grpo-opencoder-50k
This dataset was created by converting another dataset to JSONL format.
Files
dataset.jsonl: Dataset in JSONL format
Usage
from datasets import load_dataset
dataset = load_dataset("Nutanix/grpo-opencoder-50k", data_files="dataset.jsonl")
opencode_openmath_60k_c40m60
opencode_openmath_60k_c40m60
60k instruction mixture used for RELAY / Fast-dLLM-style SFT ablations:
24,000 code rows from nvidia/OpenCodeInstruct (40%)
36,000 math rows from nvidia/OpenMathInstruct-2 (train_1M, 60%)
Seed 42; code rows require a Python def; length-filtered to ≤2048 tokens under a chat template used only for filtering
Format
LMFlow-style conversation JSON (not model-specific weights or LLaDA ChatML):
{"type": "conversation", "instances":… See the full description on the dataset page: https://huggingface.co/datasets/brozonoyer/opencode_openmath_60k_c40m60.Nemotron-SFT-OpenCode-v1
Dataset Description:
Nemotron-SFT-OpenCode-v1 is an agentic instruction tuning dataset that enhances the ability of Large Language Models (LLMs) to operate within the OpenCode Command Line Interface (CLI) framework and instills simple capabilities such as tool calling and agent skills.
This dataset is ready for commercial/non-commercial use.
Dataset Subsets:
Nemotron-SFT-OpenCode-v1 contains the following subsets, where the questions and agent skills are synthetically… See the full description on the dataset page: https://huggingface.co/datasets/ashotaslanyan/Nemotron-SFT-OpenCode-v1.grpo-opencoder-small
grpo-opencoder-small
This dataset was created by converting another dataset to JSONL format.
Files
dataset.jsonl: Dataset in JSONL format
Usage
from datasets import load_dataset
dataset = load_dataset("Nutanix/grpo-opencoder-small", data_files="dataset.jsonl")
grpo-opencoder-mini
grpo-opencoder-mini
This dataset was created by converting another dataset to JSONL format.
Files
dataset.jsonl: Dataset in JSONL format
Usage
from datasets import load_dataset
dataset = load_dataset("Nutanix/grpo-opencoder-mini", data_files="dataset.jsonl")
Nemotron-SFT-OpenCode-v1
Dataset Description:
Nemotron-SFT-OpenCode-v1 is an agentic instruction tuning dataset that enhances the ability of Large Language Models (LLMs) to operate within the OpenCode Command Line Interface (CLI) framework and instills simple capabilities such as tool calling and agent skills.
This dataset is ready for commercial/non-commercial use.
Dataset Subsets:
Nemotron-SFT-OpenCode-v1 contains the following subsets, where the questions and agent skills are… See the full description on the dataset page: https://huggingface.co/datasets/Arsh9210/Nemotron-SFT-OpenCode-v1.opencode_reasoning2_hard_codeforces2000_pr03_qwen35_fp8_thinking_annotated_10k_seed20260513
Qwen3.5 FP8 Annotations for 10K K2-Think OCR2 Coding Steps
This dataset contains Qwen3.5 FP8 step-level correctness annotations for K2-Think reasoning traces on a hard Codeforces subset of OpenCodeReasoning-2.
Summary
Source trace dataset: opencode_reasoning2_hard_codeforces2000_pr03_k2_thinking_extracted_pilot10
Source rows: 10 hard coding problem traces
Candidate step rule: claim with non-empty aligned_token_ids
Candidate steps: 15,267
Manifest-selected annotated… See the full description on the dataset page: https://huggingface.co/datasets/JingweiNi/opencode_reasoning2_hard_codeforces2000_pr03_qwen35_fp8_thinking_annotated_10k_seed20260513.densemixer-ab-qwen3-30b-a3b-thinking-opencode-serveparity-idEval
DenseMixer A/B — serve-parity ID-eval traces + weight-delta/routing analysis
Full artifacts for the controlled paired-init A/B ablation testing whether DenseMixer
(training-only dense-forward + STE counterfactual router gradient; yaof20/DenseMixer,
Axolotl integrations/densemixer/) improves MoE SFT quality — the empirical answer to
marin-community/marin#7088.
Setup (identical except ONE flag)
Base / θ₀: Qwen/Qwen3-30B-A3B-Thinking-2507 @ 144afc2f (shared init).… See the full description on the dataset page: https://huggingface.co/datasets/open-athena/densemixer-ab-qwen3-30b-a3b-thinking-opencode-serveparity-idEval.