List your company
ResearchFormerly WizWand

Multimodal Understanding on MM-Bench en (test)

83.9AccuracyPVC InternVL2 (Ours)
60.91666.88372.8578.817Aug 3, 2024Aug 24, 2024Sep 15, 2024Oct 7, 2024Oct 29, 2024Nov 20, 2024Dec 12, 2024
Updated 8mo ago

Evaluation results

MethodLinks
PVC InternVL2 (Ours)Size=8B, #token /image tile=642024.1283.9
Qwen2-VLSize=7B, #token /image tile=Native resolution (unfixed)2024.1283
IXC-2.5Size=7B, #token /image tile=4002024.1282.2
InternVL2Size=8B, #token /image tile=2562024.1281.7
Oryx MLLMSize=7B2024.1281.4
LLaVA-NeXT-Yi-34BSize=34B2024.0881.1
Cambrian-34BSize=34B2024.0880.4
MiniCPM-Llama3-V 2.5Size=8.5B2024.0877.2
GPT-4V (2023.11.06)2024.0877
Bunny-Llama-3-8BSize=8.4B2024.0877
PVC InternVL2 (Ours)Size=2B, #token /image tile=642024.1275.8
Idefics2Size=8.0B2024.0875.7
Qwen2-VLSize=2B, #token /image tile=Native resolution (unfixed)2024.1274.9
Cambrian-8BSize=8.3B2024.0874.6
CogVLM2-19B-ChatSize=19B2024.0873.9
DeepSeek-VL-7BSize=7.3B2024.0873.8
Gemini Pro2024.0873.6
InternVL2Size=2B, #token /image tile=2562024.1273.2
Yi-VL-34BSize=34B2024.0872.4
XTuner-Llama-3-8B-v1.1Size=8.4B2024.0871.7
MiniCPM-V 2.0Size=2.8B2024.0869.1
Yi-VL-6BSize=6.7B2024.0868.4
DeepSeek-VL-1.3BSize=1.7B2024.0866.4
CogVLM-ChatSize=17B2024.0865.8
MiniCPM-V 1.0Size=2.8B2024.0864.1
Claude 3 Opus2024.0863.3
Qwen-VL-ChatSize=9.6B2024.0861.8