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

Math-PT (2026) tests mathematical reasoning on problems written natively in Brazilian Portuguese; this board covers the level-4 open-ended questions from OBM. Accuracy in percent, with Kimi K2 Thinking checking that the final answer is mathematically equivalent to the gold answer, higher is better.

Metric: Accuracy (%) on the 38 level-4 open-ended questions in Brazilian Portuguese (OBM); native-language zero-shot prompt asking for a boxed final answer, figures given as LaTeX source; Kimi K2 Thinking judges whether the final answer is mathematically equivalent to the gold answer; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 11 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-589.47#178
2Qwen 3 235B A22B86.84#628
3DeepSeek V3.181.58#511
4Qwen 3 14B78.95#1269
5Qwen 3 32B76.32#972
6Gemini 2.5 Flash71.05#463
7Claude Haiku 4.560.53#583
8Gemma 3 27B (IT)52.63#1324
9Qwen 3 8B52.63#1873
10Qwen 3 30B A3B50#1156
11Llama 3.3 70B Instruct18.42#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-oe-level-4 · How It Works · Data refreshed daily, snapshot 2026-11-21.