Math-PT (pt-BR MC) - Level 4: leaderboard

Math-PT (2026) tests mathematical reasoning on problems written natively in Brazilian Portuguese; this board covers the level-4 multiple-choice questions from OBMEP and OMIF. Accuracy in percent on the boxed answer letter, higher is better.

Metric: Accuracy (%) on the 270 level-4 multiple-choice questions in Brazilian Portuguese (OBMEP and OMIF); 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: Estimated already saturated. 11 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-593.33#178
2Qwen 3 235B A22B91.48#628
3DeepSeek V3.189.26#511
4Qwen 3 14B88.15#1269
5Gemini 2.5 Flash85.93#463
6Qwen 3 32B85.19#972
7Qwen 3 30B A3B84.44#1156
8Claude Haiku 4.581.48#583
9Qwen 3 8B81.48#1873
10Gemma 3 27B (IT)64.81#1324
11Llama 3.3 70B Instruct32.96#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-4 · How It Works · Data refreshed daily, snapshot 2026-11-21.