Math-PT (pt-PT MC) - Level 1: leaderboard

Math-PT (2026) tests mathematical reasoning on problems written natively in European Portuguese; this board covers the level-1 multiple-choice questions from OPM. Accuracy in percent on the boxed answer letter, higher is better.

Metric: Accuracy (%) on the 78 level-1 multiple-choice questions in European Portuguese (OPM); 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-596.15#178
2Qwen 3 235B A22B91.03#628
3Qwen 3 32B88.46#972
4Qwen 3 14B88.46#1269
5DeepSeek V3.187.18#511
6Qwen 3 8B84.62#1873
7Gemini 2.5 Flash83.33#463
8Qwen 3 30B A3B82.05#1156
9Claude Haiku 4.580.77#583
10Gemma 3 27B (IT)57.69#1324
11Llama 3.3 70B Instruct41.03#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-pt-mc-level-1 · How It Works · Data refreshed daily, snapshot 2026-11-21.