Traffic-VQA (Optical + Thermal): leaderboard
Metric: Overall accuracy (%) with paired optical and thermal infrared images, Traffic-VQA test set, zero-shot, overall accuracy over all question types; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 67.72 | #333 |
| 2 | Gemini 2.5 Flash | 67.47 | #237 |
| 3 | Gemini 2.5 Pro | 61.41 | #145 |
| 4 | Qwen 2.5 VL 7B Instruct | 43.73 | #643 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=traffic-vqa-optical-plus-thermal · How It Works · Data refreshed daily, snapshot 2026-10-11.