OpenCodeReasoning-Nemotron-32B-IOI

OpenCodeReasoning-Nemotron-32B-IOI Overview#

Description:#

OpenCodeReasoning-Nemotron-32B-IOI is a large language model (LLM) which is a derivative of Qwen2.5-32B-Instruct (AKA the reference model). It is a reasoning model that is post-trained for reasoning for code generation. The model supports a context length of 32K tokens.

This model is ready for commercial/non-commercial use.

Image: Evaluation Results

Results from OpenCodeReasoning#

Below results are the average of 64 evaluations on each benchmark.

Reproducing our results#

• Models
• Dataset
• Paper

How to use the models?#

To run inference on coding problems for IOI Benchmark:

import transformers
import torch

model_id = "nvidia/OpenCodeReasoning-Nemotron-32B-IOI"

pipeline = transformers.pipeline(
    "text-generation",
    model=model_id,
    model_kwargs={"torch_dtype": torch.bfloat16},
    device_map="auto",
)

prompt = """You are a helpful and harmless assistant. You should think step-by-step before responding to the instruction below.

Please use c++ programming language only.

You must use ```cpp for just the final solution code block with the following format:

// Your code here

{user}
"""

messages = [
    {
        "role": "user",
        "content": prompt.format(user="Write a program to calculate the sum of the first $N$ fibonacci numbers")},
]

outputs = pipeline(
    messages,
    max_new_tokens=32768,
)
print(outputs[0]["generated_text"][-1]['content'])

To run inference on coding problems for python programs:

import transformers
import torch

model_id = "nvidia/OpenCodeReasoning-Nemotron-32B"

pipeline = transformers.pipeline(
    "text-generation",
    model=model_id,
    model_kwargs={"torch_dtype": torch.bfloat16},
    device_map="auto",
)

prompt = """You are a helpful and harmless assistant. You should think step-by-step before responding to the instruction below.

Please use python programming language only.

You must use ```python for just the final solution code block with the following format:

Your code here#

{user}
"""

messages = [
    {
        "role": "user",
        "content": prompt.format(user="Write a program to calculate the sum of the first $N$ fibonacci numbers")},
]

outputs = pipeline(
    messages,
    max_new_tokens=32768,
)
print(outputs[0]["generated_text"][-1]['content'])

Citation#

If you find the data useful, please cite:

@article{ahmad2025opencodereasoning,
      title={OpenCodeReasoning: Advancing Data Distillation for Competitive Coding}, 
      author={Wasi Uddin Ahmad, Sean Narenthiran, Somshubra Majumdar, Aleksander Ficek, Siddhartha Jain, Jocelyn Huang, Vahid Noroozi, Boris Ginsburg},
      year={2025},
      eprint={2504.01943},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2504.01943}, 
}

Additional Information#

Model Architecture:#

Architecture Type: Dense decoder-only Transformer model Network Architecture: Qwen-32B-Instruct

This model was developed based on Qwen2.5-32B-Instruct and has 32B model parameters. OpenCodeReasoning-Nemotron-32B was developed based on Qwen2.5-32B-Instruct and has 32B model parameters.

Input:#

Input Type(s): Text

Input Format(s): String

Input Parameters: One-Dimensional (1D)

Other Properties Related to Input: Context length up to 32,768 tokens

Output:#

Output Type(s): Text

Output Format: String

Output Parameters: One-Dimensional (1D)

Other Properties Related to Output: Context length up to 32,768 tokens

Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA’s hardware (e.g. GPU cores) and software frameworks (e.g., CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions.

Software Integration :#

• Runtime Engine: NeMo 2.3.0
• Recommended Hardware Microarchitecture Compatibility:

NVIDIA Ampere

NVIDIA Hopper

• Preferred/Supported Operating System(s): Linux

Model Version(s):#

1.0 (4/25/2025)

OpenCodeReasoning-Nemotron-7B

OpenCodeReasoning-Nemotron-14B

OpenCodeReasoning-Nemotron-32B

OpenCodeReasoning-Nemotron-32B-IOI

Training and Evaluation Datasets:#

Training Dataset:#

The training corpus for OpenCodeReasoning-Nemotron-32B is OpenCodeReasoning dataset, which is composed of competitive programming questions and DeepSeek-R1 generated responses.

Data Collection Method: Hybrid: Automated, Human, Synthetic

Labeling Method: Hybrid: Automated, Human, Synthetic

Properties: 736k samples from OpenCodeReasoning (https://huggingface.co/datasets/nvidia/OpenCodeReasoning)

Evaluation Dataset:#

We used the datasets listed in the next section to evaluate OpenCodeReasoning-Nemotron-32B.

Data Collection Method: Hybrid: Automated, Human, Synthetic Labeling Method: Hybrid: Automated, Human, Synthetic

License/Terms of Use:#

GOVERNING TERMS: Use of this model is governed by Apache 2.0.

Deployment Geography:#

Global

Use Case:#

This model is intended for developers and researchers building LLMs.

Release Date:#

Huggingface [04/25/2025] via https://huggingface.co/nvidia/OpenCodeReasoning-Nemotron-32B/

Reference(s):#

[2504.01943] OpenCodeReasoning: Advancing Data Distillation for Competitive Coding

Inference:#

Engine: vLLM

Test Hardware NVIDIA H100-80GB

Ethical Considerations:#

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

Please report security vulnerabilities or NVIDIA AI Concerns here.