List your company

Visual Question Answering on IRIS-120K (Judge)

97.25AccuracyIRIS-4B
30.11847.546564.97582.4035Jul 5, 2026

Evaluation results

MethodLinks
IRIS-4BGroup=Ours2026.0797.25
IRIS-8BGroup=Ours2026.0797.04
IRIS-4B-TFTGroup=Ours2026.0796.51
IRIS-2BGroup=Ours2026.0796.19
IRIS-4B-SceneGroup=Ours2026.0793.12
Lingshu-7BGroup=Med2026.0784.97
Lingshu-32BGroup=Med2026.0784.76
HuluMed-14BGroup=Med2026.0783.92
HuluMed-32BGroup=Med2026.0783.7
HuluMed-7BGroup=Med2026.0783.07
Qwen3-vl-32BGroup=Gen2026.0777.57
Internvl3.5-14BGroup=Gen2026.0777.35
Internvl3.5-8BGroup=Gen2026.0772.59
MG1.5-4BGroup=Med2026.0771.01
GPT-5-miniGroup=Gen2026.0769.52
Qwen3-vl-8BGroup=Gen2026.0769.52
Qwen3-vl-2BGroup=Gen2026.0769.31
MG-4BGroup=Med2026.0766.88
HGPT-V-34BGroup=Med2026.0766.56
Qwen3-vl-4BGroup=Gen2026.0762.43
HGPT-V-7BGroup=Med2026.0732.7