LIMO: Less Is More for Reasoning 🚀

This is the updated version (v2) of the LIMO model, corresponding to the latest paper version as of July 30, 2025.

Model Information

Model	Backbone	Size
LIMO-v2	Qwen2.5-32B-Instruct	32B

Previous Version

If you need the original LIMO model (corresponding to the initial paper version), you can access it at:

LIMO v1: GAIR/LIMO

Quick Start

Our model is fine-tuned on Qwen2.5-32B-Instruct and is compatible with most mainstream frameworks like HF Transformers, VLLM, TensorRT-LLM and etc.

Using HF Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Initialize model and tokenizer
model = AutoModelForCausalLM.from_pretrained(
    "GAIR/LIMO-v2",
    torch_dtype="auto",
    trust_remote_code=True,
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("GAIR/LIMO-v2", trust_remote_code=True)

# Prepare input messages
messages = [
    {"role": "system", "content": "Please reason step by step, and put your final answer within \\boxed{}."},
    {"role": "user", "content": "What is the result of 1+1?"}
]

# Format input using chat template
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)

# Tokenize input
inputs = tokenizer(text, return_tensors="pt").to(model.device)

# Generate response
outputs = model.generate(
    **inputs,
    max_new_tokens=32768,
    temperature=0.7,
    top_p=0.95,
    do_sample=True
)

# Decode and print response
response = tokenizer.decode(outputs[0][inputs['input_ids'].shape[1]:], skip_special_tokens=True)
print(response)

Using VLLM

from vllm import LLM, SamplingParams
from transformers import AutoTokenizer

# Initialize the model
llm = LLM(
    model="GAIR/LIMO-v2",
    tensor_parallel_size=4,  # adjust based on available GPUs
    trust_remote_code=True,
    swap_space=60,
    gpu_memory_utilization=0.96,
)

# Prepare input messages
messages = [
    {"role": "system", "content": "Please reason step by step, and put your final answer within \\boxed{}."},
    {"role": "user", "content": "What is the result of 1+1?"}
]

# Setup tokenizer
tokenizer = AutoTokenizer.from_pretrained("GAIR/LIMO-v2", trust_remote_code=True)
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)

# Configure generation parameters
sampling_params = SamplingParams(
    temperature=0.7,
    max_tokens=32768,
    top_p=0.95,
)

# Generate response
output = llm.generate(text, sampling_params)
print(output[0].outputs[0].text)

Citation

@misc{ye2025limoreasoning,
      title={LIMO: Less is More for Reasoning}, 
      author={Yixin Ye and Zhen Huang and Yang Xiao and Ethan Chern and Shijie Xia and Pengfei Liu},
      year={2025},
      eprint={2502.03387},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2502.03387}, 
}

For more details and training code, please visit our GitHub repository.

GAIR
/

LIMO-v2