huggingface/open-r1

★ 26,021PythonAudience · researcherComplexity · 5/5MaintainedLicenseSetup · hard

Mindmap

mindmap
  root((Open R1))
    What it does
      Train reasoning models
      Show thinking process
      Reproduce DeepSeek-R1
    Key insight
      Step-by-step reasoning
      Improves accuracy
      Math and coding focus
    What you get
      Training code
      Datasets
      Recipes and guides
    Requirements
      GPU infrastructure
      ML expertise
      H100 GPUs recommended
    Use cases
      Build custom models
      Research reasoning AI
      Competitive benchmarks

mindmap root((Open R1)) What it does Train reasoning models Show thinking process Reproduce DeepSeek-R1 Key insight Step-by-step reasoning Improves accuracy Math and coding focus What you get Training code Datasets Recipes and guides Requirements GPU infrastructure ML expertise H100 GPUs recommended Use cases Build custom models Research reasoning AI Competitive benchmarks

Things people build with this

USE CASE 1

Train a custom reasoning AI model on your own GPU cluster to solve math and coding problems.

USE CASE 2

Reproduce DeepSeek-R1's capabilities using open-source code and published datasets.

USE CASE 3

Build competitive programming AI systems that outperform larger commercial models.

USE CASE 4

Research how step-by-step reasoning improves AI accuracy on complex tasks.

Tech stack

PythonPyTorchHugging Face TransformersCUDAGPU training

Getting it running

Difficulty · hard Time to first run · 1day+

Requires GPU with CUDA, large datasets, and significant compute resources for model training; not just inference.

Use freely for any purpose including commercial. Keep the notice and disclose changes to the patent grant.

In plain English

Open R1 is an open-source project by Hugging Face, one of the leading AI research platforms, that aims to fully reproduce DeepSeek-R1, a breakthrough reasoning AI model released by a Chinese AI lab in early 2025. DeepSeek-R1 made waves because it demonstrated exceptional reasoning capabilities (particularly in math, coding, and science problems) at a fraction of the cost of competitors like OpenAI's models. However, DeepSeek didn't release all the training details needed to reproduce it. Open R1 is the community's effort to fill those gaps. The project provides the training code, datasets, and step-by-step recipes needed to train your own version of this type of "reasoning model", an AI that shows its thinking process step by step before giving an answer, similar to how a student might show their work on a math problem. The key insight behind these models is that training them to think through problems systematically dramatically improves accuracy on difficult tasks. This is a highly technical research project aimed at AI researchers, machine learning engineers, and teams who want to train their own advanced AI models from scratch. It requires significant GPU infrastructure, the recommended setup is 8 high-end H100 GPUs, and deep familiarity with machine learning training pipelines. For context, Hugging Face has already published several companion datasets generated from this work, and a 7-billion parameter model trained using these techniques that can outperform much larger commercial models on competitive programming benchmarks. The project is actively ongoing and collaborative.

Copy-paste prompts

Prompt 1

I want to train a reasoning model like DeepSeek-R1 using Open R1. What GPU setup and training steps do I need?

Prompt 2

Show me the training code and dataset recipes in Open R1 for building a step-by-step reasoning AI.

Prompt 3

How do I use Open R1 to fine-tune a 7-billion parameter model for competitive programming benchmarks?

Prompt 4

What are the key differences between Open R1's approach and standard language model training?

Prompt 5

Walk me through the Open R1 training pipeline from data preparation to model evaluation.

Open on GitHub → Explain another repo

Generated 2026-05-18 · Model: sonnet-4-6 · Verify against the repo before relying on details.