MGSM — leaderboard
A multilingual benchmark for mathematical questions.
Metric: Score (self-reported). Source: vals.ai. Status: saturated. 66 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.5 Anthropic | 95.2 |
| 2 | Claude Opus 4.1 Anthropic | 94.44 |
| 3 | Claude Sonnet 4.5 Anthropic | 94.33 |
| 4 | GPT-5.2 x-highOpenAI | 94 |
| 5 | Gemini 3 highGoogle | 93.93 |
| 6 | Claude Opus 4 Anthropic | 93.78 |
| 7 | o4 Mini highOpenAI | 93.42 |
| 8 | Gemini 3 Flash Preview highGoogle | 93.31 |
| 9 | Claude Sonnet 4 Anthropic | 93.02 |
| 10 | Claude 3.7 Sonnet (thinking) Anthropic | 92.98 |
Interactive version: theaggregate.ai/benchmark?slug=mgsm · How the rankings work · Data refreshed daily, snapshot 2026-07-22.