MiroMind-M1

🧾 Overview

Training performance of MiroMind-M1-RL-7B on AIME24 and AIME25.

MiroMind-M1 is a fully open-source series of reasoning language models built on Qwen-2.5, focused on advancing mathematical reasoning. It is trained through supervised fine-tuning (SFT) on 719K curated problems and reinforcement learning with verifiable rewards (RLVR) on 62K challenging examples, using a context-aware multi-stage policy optimization method (CAMPO). MiroMind-M1 achieves state-of-the-art performance among open-source 7B Qwen-2.5-based models on AIME24, AIME25, and MATH500, with all models (MiroMind-M1-SFT-7B, MiroMind-M1-RL-7B, MiroMind-M1-RL-32B), data (MiroMind-M1-SFT-719K, MiroMind-M1-RL-62K), and training setups openly released.

📊 Evaluation

MiroMind-M1-SFT

Model	Initial Checkpoint	AIME24 (avg@64)	AIME25 (avg@64)	MATH500 (avg@5)
DeepSeek-R1-Distill	Qwen2.5-Math-7B	55.5	40.4†	92.8
OpenThoughts	Qwen2.5-7-Instruct	31.3	23.3	83.2
Open-R1	Qwen2.5-Math-7B-Instruct	36.7	40.0	90.6
Synthetic-1	Qwen2.5-7B-Instruct	30.0	26.6	85.6
MiMo-7B-SFT	MiMo-7B-Base	58.7	44.3	93.0
MiroMind-SFT-7B	Qwen2.5-Math-7B	60.4	45.0	94.6

† means that the score of DeepSeek-R1 on AIME25 is from our evaluation.

MiroMind-M1-RL

Model	AIME24 (avg@64)	AIME25 (avg@64)	MATH500 (avg@5)
DeepSeek-R1	79.8	70.0	–
DeepSeek-R1-0528	91.4	87.5	–
Qwen3-8B	76.0	67.3	–
DeepSeek-R1-0528-Qwen3-8B	86.0	76.3	–
MiMo-7B-RL	68.2	55.4	95.8
32B Models trained from Qwen2.5 series
DeepSeek-R1-Distill-Qwen-32B	70.8	52.1	95.8
Skywork-OR1-32B-Preview	77.1	68.2	97.5
MiroMind-M1-RL-32B	77.5	65.6	96.4
7B Models trained from Qwen2.5 series
DeepSeek-R1-Distill-Qwen-7B	55.5	39.2	–
MiroMind-M1-SFT-7B	60.4	45.0	94.6
Light-R1-7B-DS	59.1	44.3	–
Skywork-OR1-7B	72.2	54.6	–
MiroMind-M1-RL-7B	73.4	57.8	96.7

🛠 Getting Started

Installation

venv environment:

git clone https://github.com/MiroMindAsia/MiroMind-M1.git
cd MiroMind-M1

# Install Python 3.10 environment.
python3.10 -m pip install virtualenv
virtualenv -p python3.10 venv
source venv/bin/activate

# Install dependencies.
pip3 install torch==2.4.0 --index-url https://download.pytorch.org/whl/cu124
pip3 install numpy psutil ninja packaging cmake
pip3 install flash_attn==2.7.4.post1 --no-build-isolation # This may take a while...
pip3 install -e .

🏋️ Training

Multi-Node Training

Here is a quik guided to start Ray for multi-node training.

On the head node

ray stop
ray start --head --node-ip-address $HEAD_NODE_IP --num-gpus 8 --dashboard-host=0.0.0.0

On other nodes

ray stop
ray start --address="$HEAD_NODE_IP:6379" --num-gpus 8

Start Training

First, please provde the below variables:

export MODEL_PATH=YOUR_MODEL_PATH
export CKPTS_DIR=YOUR_CKPTS_DIR
export TRAIN_FILE=YOUR_TRAIN_FILE
export TEST_FILE=YOUR_TEST_FILE
export HOME=YOUR_HOME_PATH

Then run the below script to start the training:

bash m1_train_script/campo_32b.sh

⚖️ Run Evaluation

We provide ready-to-use evaluation scripts in the eval_example_script/ directory for mathematical reasoning benchmarks.

Quick Start

# Evaluate on AIME 2024
bash eval_example_script/evaluate_7b_aime24.sh

# Evaluate on AIME 2025  
bash eval_example_script/evaluate_7b_aime25.sh

# Evaluate on Math-500
bash eval_example_script/evaluate_7b_math500.sh

Supported Benchmarks

Dataset	Script	Standard Runs
AIME 2024	evaluate_7b_aime24.sh	64 runs
AIME 2025	evaluate_7b_aime25.sh	64 runs
Math-500	evaluate_7b_math500.sh	5 runs

Results

Results are saved in results/[model_name]/[dataset_name]/ with:

• average_accuracy.txt: Final accuracy score
• run[X]_inference_eval_results.csv: Detailed results

📚 Citation

Please cite our technical report if you found our work helpful:

@article{li2025miromind,
  title={MiroMind-M1: An Open-Source Advancement in Mathematical Reasoning via Context-Aware Multi-Stage Policy Optimization},
  author={Li, Xingxuan and Xiao, Yao and Ng, Dianwen and Ye, Hai and Deng, Yue and Lin, Xiang and Wang, Bin and Mo, Zhanfeng and Zhang, Chong and Zhang, Yueyi and others},
  journal={arXiv preprint arXiv:2507.14683},
  year={2025}
}

@article{zhang2025100,
  title={100 days after deepseek-r1: A survey on replication studies and more directions for reasoning language models},
  author={Zhang, Chong and Deng, Yue and Lin, Xiang and Wang, Bin and Ng, Dianwen and Ye, Hai and Li, Xingxuan and Xiao, Yao and Mo, Zhanfeng and Zhang, Qi and others},
  journal={arXiv preprint arXiv:2505.00551},
  year={2025}
}

🙏 Acknowledgement

The RL trianing is built from the wonderful verl project.