@inproceedings{melhem-mustafa-2026-optimizing,
title = "Optimizing Inference Cost for {A}rabic{--}{E}nglish Mathematical Reasoning in {LLM}s",
author = "Melhem, Sahar and
Mustafa, Ahmad M",
editor = "Abdul-Mageed, Muhammad and
Elsayed, Tamer and
Bouamor, Houda and
El-Beltagy, Samhaa and
Al-Khalifa, Hend and
Abdelali, Ahmed and
Elmadany, AbdelRahim and
El Mekki, Abdellah and
Darwish, Kareem and
Alkhereyf, Sakhar and
Luqman, Hamzah and
Khalifa, Salam and
Magdy, Samar and
Alyafeai, Zaid and
Hamed, Injy and
Zaghouani, Wajdi and
Zitouni, Imed",
booktitle = "Proceedings of The Fourth {A}rabic Natural Language Processing Conference",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.arabicnlp-main.13/",
pages = "170--178",
ISBN = "979-8-89176-473-6",
abstract = "Large language models (LLMs) can solve mathematical problems through prompting, yet not all reasoning is cost-justified when using paid APIs. We study prompt-only methods for overthinking control in the black-box setting across four datasets: Arabic GSM8K, English GSM8K, English SVAMP, and ArMATH, each evaluated on 100-question subsets divided into Easy, Medium, and Hard categories. We test Direct Answer, Chain-of-Thought (CoT), First Answer Anchoring (FAA-lite), Confidence-Triggered Escalation (CTE), Self-Consistency, Verification, Arabic prompt variations, and Language-Optimal Prompt Routing (LOPR). CTE achieves 90{\%} accuracy on Arabic GSM8K at 1.14 calls/query, 96{\%} on English GSM8K at 1.06 calls/query, and 98{\%} on SVAMP at 1.08 calls/query. Self-Consistency leads on ArMATH at 90{\%} but costs 2.00 calls/query. Verification consistently increases token cost without accuracy gains. Format-Controlled CoT is the best single-call Arabic method; Bilingual-Glossary CoT is consistently harmful. CTE calibration reveals that explicit low-confidence outputs are rare ($<$1{\%}), so CTE succeeds mainly because high-confidence predictions are accurate, not through active uncertainty detection."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="melhem-mustafa-2026-optimizing">
<titleInfo>
<title>Optimizing Inference Cost for Arabic–English Mathematical Reasoning in LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sahar</namePart>
<namePart type="family">Melhem</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ahmad</namePart>
<namePart type="given">M</namePart>
<namePart type="family">Mustafa</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 Fourth Arabic Natural Language Processing Conference</title>
</titleInfo>
<name type="personal">
<namePart type="given">Muhammad</namePart>
<namePart type="family">Abdul-Mageed</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tamer</namePart>
<namePart type="family">Elsayed</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Samhaa</namePart>
<namePart type="family">El-Beltagy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hend</namePart>
<namePart type="family">Al-Khalifa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ahmed</namePart>
<namePart type="family">Abdelali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">AbdelRahim</namePart>
<namePart type="family">Elmadany</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Abdellah</namePart>
<namePart type="family">El Mekki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kareem</namePart>
<namePart type="family">Darwish</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sakhar</namePart>
<namePart type="family">Alkhereyf</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hamzah</namePart>
<namePart type="family">Luqman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Salam</namePart>
<namePart type="family">Khalifa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Samar</namePart>
<namePart type="family">Magdy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zaid</namePart>
<namePart type="family">Alyafeai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Injy</namePart>
<namePart type="family">Hamed</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wajdi</namePart>
<namePart type="family">Zaghouani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Imed</namePart>
<namePart type="family">Zitouni</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">979-8-89176-473-6</identifier>
</relatedItem>
<abstract>Large language models (LLMs) can solve mathematical problems through prompting, yet not all reasoning is cost-justified when using paid APIs. We study prompt-only methods for overthinking control in the black-box setting across four datasets: Arabic GSM8K, English GSM8K, English SVAMP, and ArMATH, each evaluated on 100-question subsets divided into Easy, Medium, and Hard categories. We test Direct Answer, Chain-of-Thought (CoT), First Answer Anchoring (FAA-lite), Confidence-Triggered Escalation (CTE), Self-Consistency, Verification, Arabic prompt variations, and Language-Optimal Prompt Routing (LOPR). CTE achieves 90% accuracy on Arabic GSM8K at 1.14 calls/query, 96% on English GSM8K at 1.06 calls/query, and 98% on SVAMP at 1.08 calls/query. Self-Consistency leads on ArMATH at 90% but costs 2.00 calls/query. Verification consistently increases token cost without accuracy gains. Format-Controlled CoT is the best single-call Arabic method; Bilingual-Glossary CoT is consistently harmful. CTE calibration reveals that explicit low-confidence outputs are rare (<1%), so CTE succeeds mainly because high-confidence predictions are accurate, not through active uncertainty detection.</abstract>
<identifier type="citekey">melhem-mustafa-2026-optimizing</identifier>
<location>
<url>https://aclanthology.org/2026.arabicnlp-main.13/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>170</start>
<end>178</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Optimizing Inference Cost for Arabic–English Mathematical Reasoning in LLMs
%A Melhem, Sahar
%A Mustafa, Ahmad M.
%Y Abdul-Mageed, Muhammad
%Y Elsayed, Tamer
%Y Bouamor, Houda
%Y El-Beltagy, Samhaa
%Y Al-Khalifa, Hend
%Y Abdelali, Ahmed
%Y Elmadany, AbdelRahim
%Y El Mekki, Abdellah
%Y Darwish, Kareem
%Y Alkhereyf, Sakhar
%Y Luqman, Hamzah
%Y Khalifa, Salam
%Y Magdy, Samar
%Y Alyafeai, Zaid
%Y Hamed, Injy
%Y Zaghouani, Wajdi
%Y Zitouni, Imed
%S Proceedings of The Fourth Arabic Natural Language Processing Conference
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-473-6
%F melhem-mustafa-2026-optimizing
%X Large language models (LLMs) can solve mathematical problems through prompting, yet not all reasoning is cost-justified when using paid APIs. We study prompt-only methods for overthinking control in the black-box setting across four datasets: Arabic GSM8K, English GSM8K, English SVAMP, and ArMATH, each evaluated on 100-question subsets divided into Easy, Medium, and Hard categories. We test Direct Answer, Chain-of-Thought (CoT), First Answer Anchoring (FAA-lite), Confidence-Triggered Escalation (CTE), Self-Consistency, Verification, Arabic prompt variations, and Language-Optimal Prompt Routing (LOPR). CTE achieves 90% accuracy on Arabic GSM8K at 1.14 calls/query, 96% on English GSM8K at 1.06 calls/query, and 98% on SVAMP at 1.08 calls/query. Self-Consistency leads on ArMATH at 90% but costs 2.00 calls/query. Verification consistently increases token cost without accuracy gains. Format-Controlled CoT is the best single-call Arabic method; Bilingual-Glossary CoT is consistently harmful. CTE calibration reveals that explicit low-confidence outputs are rare (<1%), so CTE succeeds mainly because high-confidence predictions are accurate, not through active uncertainty detection.
%U https://aclanthology.org/2026.arabicnlp-main.13/
%P 170-178
Markdown (Informal)
[Optimizing Inference Cost for Arabic–English Mathematical Reasoning in LLMs](https://aclanthology.org/2026.arabicnlp-main.13/) (Melhem & Mustafa, ArabicNLP 2026)
ACL