@inproceedings{rajaee-etal-2026-unlocking,
title = "Unlocking Reasoning Capability on Machine Translation in Large Language Models",
author = "Rajaee, Sara and
Vincent, Sebastian and
Berard, Alexandre and
Fadaee, Marzieh and
Marchisio, Kelly and
Kocmi, Tom",
editor = "Haddow, Barry and
Kocmi, Tom and
Koehn, Philipp and
Monz, Christof",
booktitle = "Proceedings of the Eleventh Conference on Machine Translation",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.wmt-main.33/",
pages = "590--606",
ISBN = "979-8-89176-474-3",
abstract = "Reasoning-oriented large language models (RLMs) achieve strong gains on tasks such as mathematics and coding by generating explicit intermediate reasoning. However, their impact on machine translation (MT) remains underexplored. We systematically evaluate several RLMs on the WMT24++ benchmark and find that enabling explicit reasoning consistently degrades translation quality across languages and models. Our structural analysis shows that MT reasoning traces are highly linear, lacking revision, self-correction, and exploration of alternative translations, which limits their usefulness. Controlled reasoning-injection experiments demonstrate that providing high-quality reasoning traces from stronger models does not reliably improve weaker models' performance, indicating the failure lies within the format of MT reasoning. To address this mismatch, we propose a structured reasoning framework tailored to translation, based on multi-step improvements and dynamic iterative revision over difficult segments. Post-training a 111B RLM on such structured reasoning traces yields consistent gains over standard translation fine-tuning and injected generic reasoning baselines. Our findings demonstrate that reasoning must be task-structured to benefit MT."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="rajaee-etal-2026-unlocking">
<titleInfo>
<title>Unlocking Reasoning Capability on Machine Translation in Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sara</namePart>
<namePart type="family">Rajaee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sebastian</namePart>
<namePart type="family">Vincent</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alexandre</namePart>
<namePart type="family">Berard</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marzieh</namePart>
<namePart type="family">Fadaee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kelly</namePart>
<namePart type="family">Marchisio</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tom</namePart>
<namePart type="family">Kocmi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Eleventh Conference on Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Barry</namePart>
<namePart type="family">Haddow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tom</namePart>
<namePart type="family">Kocmi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Philipp</namePart>
<namePart type="family">Koehn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Christof</namePart>
<namePart type="family">Monz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-474-3</identifier>
</relatedItem>
<abstract>Reasoning-oriented large language models (RLMs) achieve strong gains on tasks such as mathematics and coding by generating explicit intermediate reasoning. However, their impact on machine translation (MT) remains underexplored. We systematically evaluate several RLMs on the WMT24++ benchmark and find that enabling explicit reasoning consistently degrades translation quality across languages and models. Our structural analysis shows that MT reasoning traces are highly linear, lacking revision, self-correction, and exploration of alternative translations, which limits their usefulness. Controlled reasoning-injection experiments demonstrate that providing high-quality reasoning traces from stronger models does not reliably improve weaker models’ performance, indicating the failure lies within the format of MT reasoning. To address this mismatch, we propose a structured reasoning framework tailored to translation, based on multi-step improvements and dynamic iterative revision over difficult segments. Post-training a 111B RLM on such structured reasoning traces yields consistent gains over standard translation fine-tuning and injected generic reasoning baselines. Our findings demonstrate that reasoning must be task-structured to benefit MT.</abstract>
<identifier type="citekey">rajaee-etal-2026-unlocking</identifier>
<location>
<url>https://aclanthology.org/2026.wmt-main.33/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>590</start>
<end>606</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Unlocking Reasoning Capability on Machine Translation in Large Language Models
%A Rajaee, Sara
%A Vincent, Sebastian
%A Berard, Alexandre
%A Fadaee, Marzieh
%A Marchisio, Kelly
%A Kocmi, Tom
%Y Haddow, Barry
%Y Kocmi, Tom
%Y Koehn, Philipp
%Y Monz, Christof
%S Proceedings of the Eleventh Conference on Machine Translation
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-474-3
%F rajaee-etal-2026-unlocking
%X Reasoning-oriented large language models (RLMs) achieve strong gains on tasks such as mathematics and coding by generating explicit intermediate reasoning. However, their impact on machine translation (MT) remains underexplored. We systematically evaluate several RLMs on the WMT24++ benchmark and find that enabling explicit reasoning consistently degrades translation quality across languages and models. Our structural analysis shows that MT reasoning traces are highly linear, lacking revision, self-correction, and exploration of alternative translations, which limits their usefulness. Controlled reasoning-injection experiments demonstrate that providing high-quality reasoning traces from stronger models does not reliably improve weaker models’ performance, indicating the failure lies within the format of MT reasoning. To address this mismatch, we propose a structured reasoning framework tailored to translation, based on multi-step improvements and dynamic iterative revision over difficult segments. Post-training a 111B RLM on such structured reasoning traces yields consistent gains over standard translation fine-tuning and injected generic reasoning baselines. Our findings demonstrate that reasoning must be task-structured to benefit MT.
%U https://aclanthology.org/2026.wmt-main.33/
%P 590-606
Markdown (Informal)
[Unlocking Reasoning Capability on Machine Translation in Large Language Models](https://aclanthology.org/2026.wmt-main.33/) (Rajaee et al., WMT 2026)
ACL