Math-PT (pt-PT OE) - Level 3: leaderboard

Math-PT (2026) tests mathematical reasoning on problems written natively in European Portuguese; this board covers the level-3 open-ended questions from OPM. 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 193 level-3 open-ended questions in European Portuguese (OPM); 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: Estimated already saturated. 11 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-592.75#178
2Qwen 3 235B A22B81.35#628
3Qwen 3 14B76.68#1269
4Gemini 2.5 Flash74.61#463
5Qwen 3 32B73.06#972
6DeepSeek V3.172.02#511
7Qwen 3 30B A3B63.73#1156
8Qwen 3 8B61.14#1873
9Claude Haiku 4.556.48#583
10Gemma 3 27B (IT)33.16#1324
11Llama 3.3 70B Instruct18.13#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-oe-level-3 · How It Works · Data refreshed daily, snapshot 2026-11-21.