Math-PT (pt-PT OE) - Level 4: leaderboard
Metric: Accuracy (%) on the 221 level-4 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: Around January 2027. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 85.52 |
| 2 | Qwen 3 235B A22B | 78.73 |
| 3 | Gemini 2.5 Flash | 71.04 |
| 4 | Qwen 3 14B | 71.04 |
| 5 | DeepSeek V3.1 | 66.52 |
| 6 | Qwen 3 32B | 65.16 |
| 7 | Qwen 3 30B A3B | 55.2 |
| 8 | Qwen 3 8B | 50.23 |
| 9 | Claude Haiku 4.5 | 47.51 |
| 10 | Gemma 3 27B (IT) | 27.6 |
| 11 | Llama 3.3 70B Instruct | 20.36 |
Interactive version: theaggregate.ai/benchmark?slug=math-pt-pt-pt-oe-level-4 · How It Works · Data refreshed daily, snapshot 2026-10-07.