List your company
ResearchFormerly WizWand

STEM & Reasoning on MMMU-Pro

81AccuracyGemini 3-Pro
18.0834.41550.7567.085Feb 4, 2026Mar 1, 2026Mar 26, 2026Apr 20, 2026May 15, 2026Jun 9, 2026Jul 5, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Gemini 3-Pro2026.0281
GPT-5tier=High2026.0278.4
Gemini 2.5-Pro2026.0268.8
ERNIE 5.02026.0268.63
Qwen3-VLmode=Thinking2026.0268.28
Qwen3-VL-8BFine-tuning setting=Base2026.0756.71
Qwen3-VL-8BFine-tuning setting=CPO2026.0753.7
Qwen3-VL-8BFine-tuning setting=LoRA2026.0741.98
Qwen3-VL-8BFine-tuning setting=GRPO2026.0740.56
Qwen3-VL-8BFine-tuning setting=GSPO2026.0731.41
InternVL2-Pt + IRABackbone=InternVL2, Training stage=IRA2026.0731.2
InternVL2.5-Pt + IRABackbone=InternVL2.5, Training stage=IRA2026.0730.6
InternVL2.5-Pt + SFTBackbone=InternVL2.5, Training stage=SFT2026.0730.4
InternVL2-Pt + SFTBackbone=InternVL2, Training stage=SFT2026.0730.1
LLaVA-OV-Pt + IRABackbone=LLaVA-OV, Training stage=IRA2026.0728
LLaVA-OV-Pt + SFTBackbone=LLaVA-OV, Training stage=SFT2026.0727.4
InternVL2.5-PtBackbone=InternVL2.5, Training stage=Pre-trained2026.0726.7
InternVL2-PtBackbone=InternVL2, Training stage=Pre-trained2026.0726
Qwen3-VL-8BFine-tuning setting=FFT2026.0724.91
LLaVA-OV-PtBackbone=LLaVA-OV, Training stage=Pre-trained2026.0720.5