K-MetBench — leaderboard
Expert meteorology benchmark over 1,774 Korean National Meteorological Engineer Examination questions, including reasoning, geo-cultural, text-only, and multimodal subsets.
Metric: Accuracy (self-reported). Source: benchmarklist.com. Status: saturation imminent. 59 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.2 | 87.8 |
| 2 | Qwen 3 VL 235B A22B (Thinking) | 84.4 |
| 3 | Qwen 3.5 27B | 83 |
| 4 | Qwen 3.6 35B A3B | 82.9 |
| 5 | Qwen 3 VL 32B (Thinking) | 78.6 |
| 6 | command-a-reasoning-08-2025 | 77.8 |
| 7 | GPT-OSS-120B | 77.3 |
| 8 | Qwen 3 30B A3B 2507 (Thinking) | 76.7 |
| 9 | EXAONE 4.5 33B | 75.9 |
| 10 | Qwen 3.5 9B | 74.9 |
| 11 | Qwen 3 VL 30B A3B (Thinking) | 74.9 |
| 12 | Qwen 3 14B | 73.7 |
| 13 | Qwen 3 VL 235B A22B Instruct | 72.4 |
| 14 | Qwen 3 VL 8B (Thinking) | 71.7 |
| 15 | GPT-OSS-20B | 71.5 |
Interactive version: theaggregate.ai/benchmark?slug=k-metbench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.