math benchmark
Multilingual MGSM (CoT)
Multilingual Grade School Math (MGSM) benchmark evaluates language models' chain-of-thought reasoning abilities across ten typologically diverse languages. Contains 250 grade-school math problems manually translated from GSM8K dataset into languages including Bengali and Swahili.
Updated Aug 11, 2026
Models3
Model coverage3
MetricScore
EvidenceB
Multilingual MGSM (CoT) Ranking
Higher score ranks better on this benchmark.
| 01 | ME | 91.6% | 100.0% | 3 | C | |
| 02 | ME | 86.9% | 50.0% | 3 | C | |
| 03 | ME | 68.9% | 0.0% | 3 | C |
Multilingual MGSM (CoT) Score Distribution
A closer view of the leading scores on this benchmark.
Multilingual MGSM (CoT) Highlights
The leading models and scores on this benchmark.
What is Multilingual MGSM (CoT)?
What Multilingual MGSM (CoT) measures and how its scores work.
Multilingual Grade School Math (MGSM) benchmark evaluates language models' chain-of-thought reasoning abilities across ten typologically diverse languages. Contains 250 grade-school math problems manually translated from GSM8K dataset into languages including Bengali and Swahili.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Multilingual MGSM (CoT)
- Modality
- text
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- multilingual-mgsm-(cot)|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about Multilingual MGSM (CoT).
Which model scores highest on Multilingual MGSM (CoT)?
Llama 3.1 405B Instruct is currently ranked first with 91.6%.
What does Multilingual MGSM (CoT) measure?
Multilingual Grade School Math (MGSM) benchmark evaluates language models' chain-of-thought reasoning abilities across ten typologically diverse languages. Contains 250 grade-school math problems manually translated from GSM8K dataset into languages including Bengali and Swahili.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
3 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.