HAERAE-Vision - Transportation: leaderboard
Metric: Checklist Score (%; original under-specified queries, GPT-5-mini judge, mean of 3 runs). Source: arxiv.org. Saturation forecast: Around December 2026. 45 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 57.69 |
| 2 | GPT-5 | 54.97 |
| 3 | Grok 4 | 50.43 |
| 4 | GPT-5 Mini | 49.22 |
| 5 | Gemini 2.5 Flash | 48.12 |
| 6 | Qwen 3 VL 30B A3B (Thinking) | 37.4 |
| 7 | Qwen 3 VL 235B A22B Instruct | 36.33 |
| 8 | Qwen 3 VL 32B Instruct | 36.02 |
| 9 | Qwen 3 VL 32B (Thinking) | 35.57 |
| 10 | Qwen 3 VL 235B A22B (Thinking) | 34.12 |
| 11 | Gemini 2.5 Flash Lite | 34.1 |
| 12 | Qwen 3 VL 30B A3B Instruct | 30.22 |
| 13 | GPT-5 Nano | 26.17 |
| 14 | Mistral Medium 3.1 | 25.46 |
| 15 | Qwen 3 VL 8B (Thinking) | 24.9 |
Interactive version: theaggregate.ai/benchmark?slug=haerae-vision-transportation · How It Works · Data refreshed daily, snapshot 2026-09-25.