datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
code_generation
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
🏠 Home Page •
💻 GitHub Repository •
🏆 Leaderboard •
LiveCodeBench is a "live" updating benchmark for holistically evaluating code related capabilities of LLMs.
Particularly, it evaluates LLMs across a range of capabilties including code generation, self-repair, test output prediction, and code execution.
This is the code generation scenario of LiveCodeBench. It is also… See the full description on the dataset page: https://huggingface.co/datasets/livecodebench/code_generation.code_generation_lite
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
🏠 Home Page •
💻 GitHub Repository •
🏆 Leaderboard •
📄 Paper
Change Log
Since LiveCodeBench is a continuously updated benchmark, we provide different versions of the dataset. Particularly, we provide the following versions of the dataset:
release_v1: The initial release of the dataset with problems released between May 2023 and Mar 2024 containing 400… See the full description on the dataset page: https://huggingface.co/datasets/lighteval/code_generation_lite.livecodebench-code_generation_litelivecodebench_code_generation_lite_parquet
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
🏠 Home Page •
💻 GitHub Repository •
🏆 Leaderboard •
📄 Paper
Change Log
Since LiveCodeBench is a continuously updated benchmark, we provide different versions of the dataset. Particularly, we provide the following versions of the dataset:
release_v1: The initial release of the dataset with problems released between May 2023 and Mar 2024 containing 400… See the full description on the dataset page: https://huggingface.co/datasets/marianna13/livecodebench_code_generation_lite_parquet.text-code-galeras-code-generation-from-docstring-3k-dedupedlivecodebench-code_generation_litelca-library-based-code-generation
🏟️ Long Code Arena (Library-based code generation)
This is the benchmark for Library-based code generation task as part of the
🏟️ Long Code Arena benchmark.
The current version includes 150 manually curated instructions asking the model to generate Python code using a particular library.
The samples come from 62 Python repositories.
All the samples in the dataset are based on reference example programs written by authors of the respective libraries.
All the repositories are… See the full description on the dataset page: https://huggingface.co/datasets/JetBrains-Research/lca-library-based-code-generation.livecodebench_code_generation_liteai-code-generation-swe-agents-2026
💻 AI Code Generation, SWE Agents & Program Synthesis Dataset (2026 Edition)
A structured research dataset featuring 3,181 domain-verified research papers and 771 official code repositories focused on Autonomous Software Engineering Agents (SWE-bench), Program Synthesis, DeepSeek-Coder-V2, Qwen2.5-Coder, Test-Driven Code Repair, Self-Healing Software, AST Semantic Modeling, and Formal Logic Verification (2023–2026).
Built with Universal Scientific Engine V17.1 Gold, providing 47… See the full description on the dataset page: https://huggingface.co/datasets/beatsprom/ai-code-generation-swe-agents-2026.java_code_api_generationlivecodebench-code_generation_liteCode-Generation-Quality-Estimation
Code Generation Quality Estimation
This repository contains model-ready task context, generated code, and complete-case execution-resource targets for five public LLM code-generation cohorts. It provides deterministic 70/10/20 group-aware split versions using seeds 42, 1234, and 2026.
Configurations
There are 15 configurations: one for each dataset and split seed. Each configuration has train, validation, and test splits.
Dataset
Complete rows
Groups
Models… See the full description on the dataset page: https://huggingface.co/datasets/Samsoup/Code-Generation-Quality-Estimation.LiveCodeBench-CodeGenerationcode_generation_lite-th
LiveCodeBench code_generation_lite, Thai
111 competitive-programming problems from LeetCode and AtCoder, with the problem
statement translated to Thai. Everything else — test cases, starter code, metadata —
is the upstream value unchanged.
Known defects
The line breaks are gone from the problem statements. 110 of the 111 rows have no
line break at all in question_content; the one remaining row has two. These are
competitive-programming statements whose input and… See the full description on the dataset page: https://huggingface.co/datasets/iapp/code_generation_lite-th.leetcode_code_generationsynthetic-code-generationsThis dataset was synthetically generated using mixtral8x7b to create unique instructions following the MagicCoder Paper and reproducing the results by modifying specific attributes (snippets are larger, instructions/responses are larger, and more specific).
Below is the prompt used to generate the instruction set:
prompt=f"""<s>[INST] You are an incredibly intelligent programming AI with expertise in CloudFormation, Terraform, AWS CDK and {lang}. Please gain inspiration from the following… See the full description on the dataset page: https://huggingface.co/datasets/VishaalY/synthetic-code-generations.nature_sound_generationA dataset from downloaded youtube videos of nature sounds. Each audio clip is two seconds long.
code-generation-sft-100k
Code Generation SFT (100K)
100,000 ShareGPT conversations covering code generation across 8 programming languages, 21 categories, and 22 distinct programming tasks. Each example includes a detailed natural language request and a complete, working implementation with explanations of key design decisions.
Motivation
Coding assistants are the highest-adoption LLM application category, but most open training datasets focus on isolated functions without context. This… See the full description on the dataset page: https://huggingface.co/datasets/stindardlogic/code-generation-sft-100k.lcb_code_generation_lite
Dataset information
Copy of livecodebench/code_generation_lite.
arch-code-transfer-lpi-260903T0846-w2-raw-generationsbenchname-library-based-code-generation
🥷 BenchName (Library-based code generation)
This is the benchmark for Library-based code generation task as part of the
🥷 BenchName benchmark.
The current version includes 150 manually curated instructions asking the model to generate Python code using a particular library.
The samples come from 62 Python repositories.
All the samples in the dataset are based on reference example programs written by authors of the respective libraries.
All the repositories are published under… See the full description on the dataset page: https://huggingface.co/datasets/anon-iclr-submission/benchname-library-based-code-generation.YAP470_Code_Generation_Datasetdeckergui-code-generation
Code generation and reasoning pairs from DeckerGUI development sessions. Contains user instructions, context, and model-generated code for TypeScript, Markdown, and JSON.
Dataset Details
Repository: ctaxnagomi/deckergui-code-generation
License: MIT
DeckerGUI Version: v2.0.0
Created: 2026-08-17
Dataset Schema
See metadata.json for the full schema definition.
Usage
from datasets import load_dataset
ds =… See the full description on the dataset page: https://huggingface.co/datasets/ctaxnagomi/deckergui-code-generation.Code-Generation-LLM-LoRAcode-text-galeras-commit-generation-3k-dedupedcode-generation-pycode_generationunit_test_generation⚠️ Note: The dataset symprompt_supp.jsonl is not created by us. We only supplemented this dataset with additional branch-level metadata (e.g., has_branch, total_branches) to enable coverage testing.
This helps users keep their workflows clean when determining whether branches exist, simplifying branch coverage calculation.
It originates from the paper:
Code-Aware Prompting: A Study of Coverage Guided Test Generation in Regression Setting using LLM
— Gabriel Ryan, Siddhartha Jain, Mingyue… See the full description on the dataset page: https://huggingface.co/datasets/Code-TREAT/unit_test_generation.code_review_generationLLM-ABAP-Code-Generation-Benchmark
LLM Benchmark ABAP Code Generation Dataset
This dataset is designed for benchmarking Large Language Models (LLMs) on ABAP code generation capabilities. It is based on the HumanEval benchmark, adapted for ABAP, and includes 16 additional ABAP-specific tasks that require interaction with database tables.
Total tasks: 180
164 tasks adapted from HumanEval
16 ABAP-specific tasks
Dataset Structure
dataset.jsonl: Contains 180 examples. Each example has:
id: Unique… See the full description on the dataset page: https://huggingface.co/datasets/timkoehne/LLM-ABAP-Code-Generation-Benchmark.