Image Captioning on CapMAS (test)
89.99CLAIRVCap (e2)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VCap (e2)Iteration=e22026.05 | 89.99 | 74.18 | 86.42 | 83.53 | |
| Gemini 3.1 Pro2026.05 | 89.19 | 72 | 78.37 | 79.85 | |
| VCap (e1)Iteration=e12026.05 | 89.14 | 73.95 | 81.48 | 81.52 | |
| GPT-5.42026.05 | 88.18 | 73.92 | 81.99 | 81.36 | |
| Seed 2.0 Pro2026.05 | 88.03 | 73.43 | 77.68 | 79.71 | |
| Qwen3-VL-8B-InstrModel size=8B, Type=Instruction-tuned2026.05 | 86.19 | 71.3 | 76.25 | 77.91 | |
| Qwen3.5-397BModel size=397B2026.05 | 84.97 | 73.18 | 77.02 | 78.39 | |
| Qwen3-VL-235B-InstrModel size=235B, Type=Instruction-tuned2026.05 | 84.44 | 71.23 | 77.12 | 77.6 | |
| CapRL2026.05 | 79.65 | 69.53 | 63.76 | 70.98 |