Remote Sensing Vision-Language Perception and Reasoning on GeoReason-Bench (test)
34.92Count AccuracyGeoReason
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| GeoReasonModel Category=Open-source Remote Sensing Vision-Language Models2026.01 | 34.92 | 56.7 | 50.62 | 95.23 | 43.51 | 51.27 | 56.2 | |
| Qwen2.5-VLModel Category=Open-source Vision-Language Models2026.01 | 25.4 | 40.21 | 37.04 | 90.48 | 23.86 | 35.65 | 43.4 | |
| RS-EoTModel Category=Open-source Remote Sensing Vision-Language Models2026.01 | 19.05 | 45.36 | 45.68 | 68.25 | 32.28 | 38.71 | 42.12 | |
| SkySenseGPTModel Category=Open-source Remote Sensing Vision-Language Models2026.01 | 14.29 | 24.74 | 23.46 | 93.65 | 14.39 | 25.81 | 34.1 | |
| GeoChatModel Category=Open-source Remote Sensing Vision-Language Models2026.01 | 12.7 | 17.53 | 20.99 | 90.48 | 16.49 | 24.79 | 31.64 | |
| LlavaModel Category=Open-source Vision-Language Models2026.01 | 9.52 | 44.33 | 18.52 | 90.48 | 16.84 | 28.69 | 35.94 | |
| GPT-4oModel Category=Close-source Commercial Vision-Language Models2026.01 | 6.35 | 42.27 | 34.57 | 92.06 | 33.68 | 38.54 | 41.79 |