Math-PT (pt-BR MC) - Level 5: leaderboard

Math-PT (2026) tests mathematical reasoning on problems written natively in Brazilian Portuguese; this board covers the level-5 multiple-choice questions from ELLM and ITA. Accuracy in percent on the boxed answer letter, higher is better.

Metric: Accuracy (%) on the 110 level-5 multiple-choice questions in Brazilian Portuguese (ELLM and ITA); native-language zero-shot prompt asking for a boxed final answer, figures given as LaTeX source; the letter in the final boxed answer must match the key; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 11 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3 14B90.91#1269
2GPT-590#178
3Gemini 2.5 Flash89.09#463
4Qwen 3 235B A22B89.09#628
5Qwen 3 32B87.27#972
6DeepSeek V3.186.36#511
7Claude Haiku 4.581.82#583
8Qwen 3 30B A3B80#1156
9Qwen 3 8B76.36#1873
10Gemma 3 27B (IT)61.82#1324
11Llama 3.3 70B Instruct31.82#1385

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=math-pt-pt-br-mc-level-5 · How It Works · Data refreshed daily, snapshot 2026-11-21.