The study investigates the performance gap between native-language reasoning and English-pivoted reasoning in large language models. By creating extensive multilingual reasoning datasets and fine‑tuning specialists on Qwen/Qwen3-8B-Base, the authors find that the native reasoning gap is much smaller (1.9–3.5%) than previously reported. They analyze weight‑space changes, discover a language‑agnostic reasoning core in the middle layers, and propose a Layer Swap technique that transfers these mid‑layer updates from an English specialist to native specialists, effectively closing most of the gap while maintaining native chain‑of‑thought output.
By Maxence Lasbordes, Am\'elie Chatelain, Djam\'e Seddah
arXiv:2606. 15080v1 Announce Type: cross Abstract: While Large Reasoning Models (LRMs) show strong performance in English, they often fail to reason in the language of the query, a phenomenon known as language collapse.
By Dayeon Ki, Kevin Duh, Marine Carpuat
arXiv:2608. 10444v1 Announce Type: cross Abstract: Large language models (LLMs) have made substantial progress on reasoning tasks that require increasingly long and complex inferential chains.
By Si'an Xie (Beijing University of Posts and Telecommunications), Jiaxun Liu (Peking University), Biao Yang (Kuaishou Technology), Wei Yuan (Kuaishou Technology), Fan Yang (Kuaishou Technology), Tingting Gao (Kuaishou Technology), Ming Wu (Beijing University of Posts and Telecommunications)
arXiv:2609.10445v1 Announce Type: new
Abstract: Reasoning language models have made substantial advances on a variety of complex tasks, yet their capabilities remain overwhelmingly English-centric: m...
By Mehrnaz Mofakhami, Ananya Sahu, Alejandro R. Salamanca, Daniel D'souza, Alexandre Berard, Thomas Euyang, Marzieh Fadaee, Julia Kreutzer
arXiv:2609.37104v1 Announce Type: cross
Abstract: Open-source reasoning models provide unequal access to reasoning capability across languages. When a model can solve a problem but cannot deliver a c...
By Hongyang Li, Xiao Li, Caesar Wu, Gr\'egoire Danoy, Pascal Bouvry
The paper investigates how mathematical reasoning improvements from Reinforcement Post-Training (RPT) in English-centric large language models (LLMs) transfer to other languages. Experiments show that transfer varies greatly depending on the initial model, target language, and training approach, and that stronger English initialization does not guarantee better transfer efficiency. Adding a single parallel language to the training data yields the largest boost, with additional languages providing smaller but consistent gains.
By Wen Yang, Junhong Wu, Chong Li, Chengqing Zong, Jiajun Zhang
The study explores how the language used for reasoning affects retrieval‑augmented generation (RAG) in a monolingual German setting. Using a German RAG question‑answering testbed based on the tabletop game The Dark Eye, the authors show that aligning the reasoning language with the query and retrieved documents improves performance, with German reasoning outperforming French reasoning. However, German reasoning still does not surpass the model’s native English reasoning, indicating that native multilingual reasoning is necessary for optimal results.
By Oliver Hauck, Mario Sanz-Guerrero, Katharina von der Wense
arXiv:2608. 13698v1 Announce Type: cross Abstract: Reinforcement Learning with Verifiable Rewards (RLVR), often optimized with Group Relative Policy Optimization (GRPO), has become a central recipe for improving the reasoning capabilities of pretrained language models but current studies remain heavily English-centric.
By Konstantin Dobler, Federico Scozzafava, Jonathan Janke, Mohamed Ali, Simon Lehnerer
arXiv:2607. 25069v1 Announce Type: cross Abstract: Automated verification of numerical claims is a challenging problem, as it requires both language understanding and quantitative reasoning.
By Sagnik Sinha, Shreyas Shrestha
The paper introduces a trace-level divergence framework for evaluating multilingual large language models, breaking down agreement into structural similarity, order consistency, and holistic similarity. Experiments across three reasoning benchmarks, seven languages, and six leading LLMs show that conventional holistic similarity metrics overstate agreement, with up to 67.3% of top-scoring examples lacking structural alignment. Human judgments align more closely with structural similarity, highlighting language-specific divergence and the need for trace-level evaluation in multilingual deployments.
By Rui Sun, Peizhao Mei, Xiewei Cheng, Kaizhuo Chen, Taoran Li, Zhiyuan Yu, Varun Chandrasekaran
arXiv:2607. 05992v1 Announce Type: cross Abstract: Mathematical reasoning has become a central task for evaluating and tuning reasoning Large Language Models (LLMs), yet existing benchmarks remain heavily biased toward high-resource languages, with English and Chinese dominating both pre-training corpora and evaluation suites.
By Daryna Dementieva, Nikolay Babakov, Kathy H\"ammerl, Ilseyar Alimova, Jind\v{r}ich Libovick\'y, Shu Okabe, Miras Baisbay, Lukas Edman, Abrorkhon Inomkhujaev, Antonia Karamolegkou, Mateusz Lango, Volkan \"Ozer, Nikola Selic, Subhankar Swain, Tsedeniya Kinfe Temesgen, Galit Bary Weisberg, Alexander Fraser
Reasoning language models (RLMs) achieve strong performance on complex reasoning tasks, but still exhibit substantial multilingual reasoning gaps, largely due to language-understanding failures in non-English inputs. English translation can mitigate these failures by expressing non-English inputs in a form that RLMs can more reliably interpret, yet translating every input is unnecessary when the model can reason reliably from the original query.