Image Captioning on RSICD
60BLEU-4SkyEyeGPT
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| SkyEyeGPTVisual Encoder=EVA-G, Text Decoder=LlamaV2-Chat2025.12 | 60 | 0.837 | 0.354 | 62.6 | 86.7 | 76.7 | 67.3 | — | |
| SkyEyeGPTTotal Params=7B2025.12 | 60 | 0.837 | — | — | 86.7 | — | — | — | |
| SkyEyeGPT2025.11 | 59.99 | 83.65 | 35.35 | — | — | — | — | — | |
| OursVisual Encoder=CLIP-Cap-4, Text Decoder=LlamaV22025.12 | 50.5 | 2.781 | 0.392 | 67.2 | 77.2 | 66.2 | 57.8 | 0.501 | |
| BITATotal Params=3B2025.12 | 50.4 | 3.054 | — | — | 77.4 | — | — | — | |
| Def. Trans.Total Params=90M*2025.12 | 49.2 | 2.581 | — | — | 75.8 | — | — | — | |
| MLCA-NETVisual Encoder=VGG16, Text Decoder=LSTM2025.12 | 46.1 | 2.356 | 0.351 | 64.6 | 75.7 | 63.4 | 53.9 | 0.444 | |
| MLCA-NETTotal Params=138M*2025.12 | 46.1 | 2.356 | — | — | 75.7 | — | — | — | |
| RS-CapRetVisual Encoder=CLIP-Cap-4, Text Decoder=LlamaV2, Finetuned=false2025.12 | 45.5 | 2.605 | 0.376 | 64.9 | 74.1 | 62.2 | 52.9 | 0.484 | |
| RS-CapRetTotal Params=7B2025.12 | 45.5 | 2.605 | — | — | 74.1 | — | — | — | |
| RS-CapRetVisual Encoder=CLIP-Cap-4, Text Decoder=LlamaV2, Finetuned=true2025.12 | 43.3 | 2.502 | 0.37 | 63.3 | 72 | 59.9 | 50.6 | 0.474 | |
| MC-NetTotal Params=138M*2025.12 | 43.3 | 2.454 | — | — | 72.8 | — | — | — | |
| SkySenseGPTModel Category=Open-source Remote Sensing Vision-Language Models2026.03 | 42.47 | — | — | — | — | — | — | — | |
| RSGPTVisual Encoder=EVA-G, Text Decoder=Vicuna2025.12 | 36.8 | 1.029 | 0.301 | 53.3 | 70.3 | 54.2 | 44 | — | |
| RSGPTTotal Params=13B2025.12 | 36.8 | 1.029 | — | — | 70.3 | — | — | — | |
| GeoMELTTotal Params=271M2025.12 | 36.5 | 2.652 | — | — | 61 | — | — | — | |
| GeoSolverModel Category=Ours2026.03 | 36.18 | — | — | — | — | — | — | — | |
| GeoZero2025.11 | 29.17 | 97.45 | 48.18 | — | — | — | — | — | |
| EarthDialCategory=Open-source Remote Sensing Vision-Language Models2025.12 | 29.09 | 85.82 | 25.2 | 24.19 | — | — | — | — | |
| EarthDialModel Category=Open-source Remote Sensing Vision-Language Models2026.03 | 29.09 | — | — | — | — | — | — | — | |
| VHMCategory=Open-source Remote Sensing Vision-Language Models2025.12 | 25.66 | 49.8 | 17.63 | 20.5 | — | — | — | — | |
| VHMModel Category=Open-source Remote Sensing Vision-Language Models2026.03 | 25.66 | — | — | — | — | — | — | — | |
| MF-RSVLMCategory=Open-source Remote Sensing Vision-Language Models2025.12 | 24.97 | 37.46 | 50.57 | 38.1 | — | — | — | — | |
| SkySenseGPTCategory=Open-source Remote Sensing Vision-Language Models2025.12 | 23.33 | 40.48 | 14.02 | 18.01 | — | — | — | — | |
| RS-EoTModel Category=Open-source Reasoning Vision-Language Models2026.03 | 23.02 | — | — | — | — | — | — | — | |
| Qwen2.5-VLCategory=Open-source Vision-Language Models2025.12 | 17.8 | 32.19 | 13.72 | 14.62 | — | — | — | — | |
| Qwen2.5-VLModel Category=Open-source Vision-Language Models2026.03 | 17.8 | — | — | — | — | — | — | — | |
| GPT-5Category=Closed-source Commercial Vision-Language Models2025.12 | 16.83 | 34.39 | 16.73 | 15.86 | — | — | — | — | |
| ChatGPT-5Model Category=Closed-source Commercial Vision-Language Models2026.03 | 16.83 | — | — | — | — | — | — | — | |
| Kimi-VL-ThinkingModel Category=Open-source Reasoning Vision-Language Models2026.03 | 15.6 | — | — | — | — | — | — | — | |
| MiniGPT-v2Category=Open-source Vision-Language Models2025.12 | 15.4 | 26.63 | 12.36 | 12.21 | — | — | — | — | |
| MiniGPT-v2Model Category=Open-source Vision-Language Models2026.03 | 15.4 | — | — | — | — | — | — | — | |
| GLM-4.1V-ThinkingModel Category=Open-source Reasoning Vision-Language Models2026.03 | 12.57 | — | — | — | — | — | — | — | |
| Claude-sonnet-4Category=Closed-source Commercial Vision-Language Models2025.12 | 11.58 | 24.57 | 13.9 | 10.63 | — | — | — | — | |
| Claude-sonnet-4Model Category=Closed-source Commercial Vision-Language Models2026.03 | 11.58 | — | — | — | — | — | — | — | |
| Gemini-2.0Category=Closed-source Commercial Vision-Language Models2025.12 | 10.85 | 21.53 | 8.71 | 9.41 | — | — | — | — | |
| Gemini-2.0-flashModel Category=Closed-source Commercial Vision-Language Models2026.03 | 10.85 | — | — | — | — | — | — | — | |
| GeoChat2025.11 | — | 12.39 | 13.48 | — | — | — | — | — |