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 (imported). Source: github.com. Source: benchmarklist.com. Saturation forecast: Around December 2026. 59 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 (Thinking) | 87.8 | #110 (GPT-5.2) |
| 2 | Qwen 3 VL 235B A22B (Thinking) | 84.4 | #238 (Qwen 3 VL 235B A22B) |
| 3 | Qwen 3.5 27B (Thinking) | 83 | #187 (Qwen 3.5 27B) |
| 4 | Qwen 3.6 35B A3B (Thinking) | 82.9 | #226 (Qwen 3.6 35B A3B) |
| 5 | Qwen 3.5 35B A3B (Thinking) | 81.8 | #256 (Qwen 3.5 35B A3B) |
| 6 | Qwen 3 VL 32B (Thinking) | 78.6 | #294 (Qwen 3 VL 32B) |
| 7 | command-a-reasoning-08-2025 | 77.8 | #329 |
| 8 | GPT-OSS-120B | 77.3 | #341 |
| 9 | Qwen 3 30B A3B 2507 (Thinking) | 76.7 | #376 (Qwen 3 30B A3B 2507) |
| 10 | A.X 4.0 | 76.1 | #424 |
| 11 | EXAONE 4.5 33B | 75.9 | #396 |
| 12 | Qwen 3 VL 30B A3B (Thinking) | 74.9 | #351 (Qwen 3 VL 30B A3B) |
| 13 | Qwen 3.5 9B (Thinking) | 74.9 | #379 (Qwen 3.5 9B) |
| 14 | Qwen 3 14B | 73.7 | #537 |
| 15 | Qwen 3 VL 235B A22B Instruct | 72.4 | #273 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=k-metbench · How It Works · Data refreshed daily, snapshot 2026-10-20.