Image Captioning on Captioning Suite
85NoCaps ScoreGenLIP
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GenLIPArch=g/16, Data=8.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 85 | 75.5 | 144.8 | |
| SigLIP2Arch=So/16, Data=40.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 84.1 | 76.4 | 139.3 | |
| GenLIPArch=So/16, Data=8.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 83.1 | 74.8 | 142.1 | |
| SigLIP2Arch=g/16, Data=40.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 82 | 72.3 | 142.7 | |
| SigLIPArch=L/16, Data=40.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 81.5 | 72 | 131.1 | |
| GenLIPArch=L/16, Data=8.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 81.1 | 71.3 | 139.4 | |
| OVision2Arch=L/16, Data=12.8B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 79.4 | 69.6 | 133.8 | |
| AIMv2Arch=L/14, Data=12.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 66.5 | 55.9 | 130.5 | |
| CLIPArch=L/14, Data=12.8B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 63.1 | 54.4 | 127.9 |