MGSM
Progress Over Time
Interactive timeline showing model performance evolution on MGSM
MGSM Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  Meta | 400B | — | — | ||
| 2 |  OpenAI | — | — | — | ||
| 3 |  Anthropic | — | — | — | ||
| 3 | Anthropic | — | — | — | ||
| 5 | 70B | — | — | |||
| 6 | OpenAI | — | — | — | ||
| 7 | Anthropic | — | — | — | ||
| 8 | Meta | 109B | — | — | ||
| 9 | OpenAI | — | 128K | $2.50 / $10.00 | ||
| 10 | OpenAI | — | — | — | ||
| 11 | OpenAI | — | 128K | $10.00 / $30.00 | ||
| 12 |  Google | — | — | — | ||
| 13 | OpenAI | — | — | — | ||
| 14 | 90B | — | — | |||
| 15 | Anthropic | — | — | — | ||
| 16 |  Alibaba Cloud / Qwen Team | 235B | — | — | ||
| 17 | Anthropic | — | — | — | ||
| 18 | Google | — | — | — | ||
| 19 |  Microsoft | 15B | — | — | ||
| 20 | Anthropic | — | — | — | ||
| 21 | OpenAI | — | — | — | ||
| 22 | 11B | — | — | |||
| 23 | Google | 8B | — | — | ||
| 24 | Microsoft | 4B | — | — | ||
| 25 | 2B | — | — | |||
| 26 | Microsoft | 60B | — | — | ||
| 27 | 3B | — | — | |||
| 28 | OpenAI | — | 16K | $0.50 / $1.50 | ||
| 29 | 2B | — | — | |||
| 29 | Google | 8B | — | — | ||
| 31 | Microsoft | 4B | — | — |
What is MGSM?
MGSM (Multilingual Grade School Math) is a benchmark of grade-school math problems. Contains 250 grade-school math problems manually translated from the GSM8K dataset into ten typologically diverse languages: Spanish, French, German, Russian, Chinese, Japanese, Thai, Swahili, Bengali, and Telugu. Evaluates multilingual mathematical reasoning capabilities.
MGSM is a text benchmark evaluating models on math and reasoning tasks. LLM Stats tracks 31 models on this benchmark, scored on a 0–1 scale. The current average is 0.8, with the leader at 0.9.
Compare leaders on the best AI for math and best AI for reasoning leaderboards.
Current leaders
Llama 4 Maverick from Meta currently leads the MGSM leaderboard with a score of 0.923 across 31 evaluated AI models.
Source paper
- Title
- Language Models are Multilingual Chain-of-Thought Reasoners
- Authors
- Freda Shi, Mirac Suzgun, Markus Freitag, Xuezhi Wang, and 8 others
- Published
- arXiv
- 2210.03057
Abstract
We evaluate the reasoning abilities of large language models in multilingual settings. We introduce the Multilingual Grade School Math (MGSM) benchmark, by manually translating 250 grade-school math problems from the GSM8K dataset (Cobbe et al., 2021) into ten typologically diverse languages. We find that the ability to solve MGSM problems via chain-of-thought prompting emerges with increasing model scale, and that models have strikingly strong multilingual reasoning abilities, even in underrepresented languages such as Bengali and Swahili. Finally, we show that the multilingual reasoning abilities of language models extend to other tasks such as commonsense reasoning and word-in-context semantic judgment. The MGSM benchmark is publicly available at https://github.com/google-research/url-nlp.
FAQ
Common questions about the MGSM benchmark and leaderboard.