List your company
ResearchFormerly WizWand

Image Captioning on MS COCO IR

68.44ScoreClean
46.464852.169957.87563.5801May 21, 2026
Updated 5mo ago

Evaluation results

MethodLinks
CleanImage Encoder=EVA-CLIP ViT-G/14, Model=InstructBLIP FlanT5XL (4.1B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0568.44—
CleanImage Encoder=OpenAI CLIP ViT-L/14, Model=LLaVA-1.6 (7B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0568—
CleanImage Encoder=OpenAI CLIP ViT-L/14, Model=OpenFlamingo (3B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0567.36—
Unified PatchImage Encoder=EVA-CLIP ViT-G/14, Model=InstructBLIP FlanT5XL (4.1B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0565.732.71
CleanImage Encoder=OpenAI CLIP ViT-L/14, Model=LLaVA-1.5 (7B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0565.4—
CleanImage Encoder=EVA-CLIP ViT-G/14, Model=Blip-2 FlanT5XL (4.1B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0565.03—
CleanImage Encoder=OpenAI CLIP ViT-L/14, Model=Blip-2 FlanT5XL ViT-L (3.4B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0562.19—
Unified PatchImage Encoder=OpenAI CLIP ViT-L/14, Model=OpenFlamingo (3B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0561.65.76
TOUAPImage Encoder=EVA-CLIP ViT-G/14, Model=InstructBLIP FlanT5XL (4.1B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0559.379.07
Unified PatchImage Encoder=OpenAI CLIP ViT-L/14, Model=LLaVA-1.5 (7B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0558.876.53
Unified PatchImage Encoder=EVA-CLIP ViT-G/14, Model=Blip-2 FlanT5XL (4.1B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0557.387.65
Unified PatchImage Encoder=OpenAI CLIP ViT-L/14, Model=LLaVA-1.6 (7B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0557.1710.83
Unified PatchImage Encoder=OpenAI CLIP ViT-L/14, Model=Blip-2 FlanT5XL ViT-L (3.4B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0557.095.1
TOUAPImage Encoder=OpenAI CLIP ViT-L/14, Model=OpenFlamingo (3B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0556.9410.42
TOUAPImage Encoder=EVA-CLIP ViT-G/14, Model=Blip-2 FlanT5XL (4.1B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0556.838.2
TOUAPImage Encoder=OpenAI CLIP ViT-L/14, Model=LLaVA-1.6 (7B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0556.4511.55
TOUAPImage Encoder=OpenAI CLIP ViT-L/14, Model=LLaVA-1.5 (7B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0555.1510.25
TOUAPImage Encoder=OpenAI CLIP ViT-L/14, Model=Blip-2 FlanT5XL ViT-L (3.4B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0554.267.93
CFGPatchImage Encoder=EVA-CLIP ViT-G/14, Model=InstructBLIP FlanT5XL (4.1B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0552.9115.53
CFGPatchImage Encoder=OpenAI CLIP ViT-L/14, Model=LLaVA-1.6 (7B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0551.5416.46
CFGPatchImage Encoder=OpenAI CLIP ViT-L/14, Model=LLaVA-1.5 (7B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0550.2515.15
CFGPatchImage Encoder=OpenAI CLIP ViT-L/14, Model=Blip-2 FlanT5XL ViT-L (3.4B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0549.1213.07
CFGPatchImage Encoder=EVA-CLIP ViT-G/14, Model=Blip-2 FlanT5XL (4.1B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0548.9716.06
CFGPatchImage Encoder=OpenAI CLIP ViT-L/14, Model=OpenFlamingo (3B), Evaluation Protocol=LLM-as-Judge (GPT-5)2026.0547.3120.05