[Submitted on 14 Apr 2026 (v1), last revised 10 Oct 2026 (this version, v2)]
Abstract:Despite advances in multilingual capabilities, most large language models (LLMs) remain English-centric in their training and, crucially, in the reasoning traces they produce: even for non-English problems, they predominantly reason in English, leaving non-English users with intermediate reasoning they cannot follow. We address this disparity with two contributions. **(a)** We introduce $texttt{ReasonXL}$, the first large-scale parallel corpus of cross-domain reasoning traces in five European languages (English, German, French, Italian, and Spanish): over 2.5M aligned samples (11B tokens) per language, each comprising prompt, reasoning trace, and final output, enabling direct supervision of language-specific reasoning. **(b)** Using $texttt{ReasonXL}$, we demonstrate that LLMs can be adapted to reason entirely in a target language through a simple two-stage pipeline of supervised fine-tuning (SFT) and reinforcement learning with verifiable rewards (RLVR), matching or exceeding the base models’ English-reasoning performance in two of four languages and recovering most of it in the remaining two.
Submission history
From: Daniil Gurgurov [view email]
[v1] Tue, 14 Apr 2026 07:10:46 UTC (239 KB)
[v2] Sat, 10 Oct 2026 06:58:01 UTC (109 KB)
