--- license: apache-2.0 language: - en - zh - es - ar - vi - ja - ko - fr - pt - th tags: - O1-like model - Math pipeline_tag: text-generation --- This repository contains the resources for our **paper** [Think Natively: Unlocking Multilingual Reasoning with Consistency-Enhanced Reinforcement Learning](https://arxiv.org/pdf/2510.07300) Large Reasoning Models (LRMs) have achieved remarkable performance on complex reasoning tasks by adopting the "think-then-answer" paradigm, which enhances both accuracy and interpretability. However, current LRMs exhibit two critical limitations when processing non-English languages: (1) They often struggle to maintain input-output language consistency; (2) They generally perform poorly with wrong reasoning paths and lower answer accuracy compared to English. These limitations significantly degrade the user experience for non-English speakers and hinder the global deployment of LRMs. To address these limitations, we propose M-Thinker, which is trained by the GRPO algorithm that involves a **Language Consistency (LC) reward** and a novel **Cross-lingual Thinking Alignment (CTA) reward**. Specifically, the LC reward defines a strict constraint on the language consistency between the input, thought, and answer. Besides, the CTA reward compares the model's non-English reasoning paths with its English reasoning path to transfer its own reasoning capability from English to non-English languages. Through an iterative RL procedure, our M-Thinker-1.5B/7B models not only achieve nearly 100% language consistency and superior performance on two multilingual benchmarks (MMATH and PolyMath), but also exhibit excellent generalization on out-of-domain languages. | Model Access | Backbone | Training data Access | | :-- | :-- | :-- | M-Thinker-7B-Iter2 (👍👍)   | M-Thinker-7B-Iter1 | [M-Thinker-7B-RL-Iter2-data](https://huggingface.co/datasets/XueZhang-bjtu/M-Thinker-7B-RL-Iter2-data) M-Thinker-7B-Iter1 (👍) | [7B-cold-start-SFT](https://huggingface.co/XueZhang-bjtu/7B-cold-start-SFT) | [M-Thinker-7B-RL-Iter1-data](https://huggingface.co/datasets/XueZhang-bjtu/M-Thinker-7B-RL-Iter1-data) [7B-cold-start-SFT](https://huggingface.co/XueZhang-bjtu/7B-cold-start-SFT) | [DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B)   | [M-Thinker-SFT-data](https://huggingface.co/datasets/XueZhang-bjtu/M-Thinker-SFT-data) M-Thinker-1.5B-Iter2 (👍👍) | M-Thinker-1.5B-Iter1 | [M-Thinker-1.5B-RL-Iter2-data](https://huggingface.co/datasets/XueZhang-bjtu/M-Thinker-1.5B-RL-Iter2-data) M-Thinker-1.5B-Iter1 (👍) | [1.5B-cold-start-SFT](https://huggingface.co/XueZhang-bjtu/1.5B-cold-start-SFT) | [M-Thinker-1.5B-RL-Iter1-data](https://huggingface.co/datasets/XueZhang-bjtu/M-Thinker-1.5B-RL-Iter1-data) [1.5B-cold-start-SFT](https://huggingface.co/XueZhang-bjtu/1.5B-cold-start-SFT) | [DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B) | [M-Thinker-SFT-data](https://huggingface.co/datasets/XueZhang-bjtu/M-Thinker-SFT-data) If you find this work useful, please consider citing our paper: ``` @misc{zhang2025thinknativelyunlockingmultilingual, title={Think Natively: Unlocking Multilingual Reasoning with Consistency-Enhanced Reinforcement Learning}, author={Xue Zhang and Yunlong Liang and Fandong Meng and Songming Zhang and Kaiyu Huang and Yufeng Chen and Jinan Xu and Jie Zhou}, year={2025}, eprint={2510.07300}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2510.07300}, } ```