List your company
ResearchFormerly WizWand

Multi-discipline Multimodal Reasoning on MMMU

61.3AccuracySophiaVL-R1-7B
31.55639.2784754.722May 22, 2025Jul 11, 2025Aug 31, 2025Oct 21, 2025Dec 10, 2025Jan 30, 2026Mar 22, 2026
Updated 7mo ago

Evaluation results

MethodLinks
SophiaVL-R1-7BParameters=7B2025.0561.3
Qwen2.5-VL-7B-InstructParameters=7B, Training Strategy=SFT+GRPO2025.0559.1
Qwen2.5-VL-7B-InstructParameters=7B, Training Strategy=Direct2025.0558.7
Qwen2.5-VL-7B-InstructParameters=7B, Training Strategy=GRPO2025.0558
LLaVA-OneVision-72BParameters=72B2025.0556.8
GPT-4V2025.0556.8
InternVL2.5-8B-VisualPRMParameters=8B2025.0556.2
R1-Onevision-7BParameters=7B2025.0551.6
Cambrian-1-34BParameters=34B2025.0549.7
LLaVA-OneVision-7BParameters=7B2025.0548.8
URSA-8BParameters=8B2025.0543.1
VanillaBackbone=LLaVA-1.5-13B, Vision token count=1442026.0336.1
QMoPBackbone=LLaVA-1.5-13B, Vision token count=1442026.0336.1
LDP-v2Backbone=LLaVA-1.5-13B, Vision token count=1442026.0335.6
C-AbstractorBackbone=LLaVA-1.5-13B, Vision token count=1442026.0335.4
QMoPBackbone=LLaVA-1.5-7B, Vision token count=1442026.0335.1
TokenPackerBackbone=LLaVA-1.5-13B, Vision token count=1442026.0335
FasterVLMBackbone=LLaVA-1.5-13B, Vision token count=1442026.0334.9
FasterVLMBackbone=LLaVA-1.5-7B, Vision token count=1442026.0334.8
MQT-LLaVABackbone=LLaVA-1.5-7B, Vision token count=1442026.0334.8
VanillaBackbone=LLaVA-1.5-7B, Vision token count=1442026.0334.6
Pixel-ShuffleBackbone=LLaVA-1.5-7B, Vision token count=642026.0334.6
FasterVLMBackbone=LLaVA-1.5-7B, Vision token count=642026.0334.6
QMoPBackbone=LLaVA-1.5-7B, Vision token count=642026.0334.6
MQT-LLaVABackbone=LLaVA-1.5-7B, Vision token count=642026.0334.4
Pixel-ShuffleBackbone=LLaVA-1.5-13B, Vision token count=1442026.0334.4
TokenPackerBackbone=LLaVA-1.5-7B, Vision token count=1442026.0334.3
LDP-v2Backbone=LLaVA-1.5-7B, Vision token count=642026.0334.2
C-AbstractorBackbone=LLaVA-1.5-7B, Vision token count=642026.0334.1
LDP-v2Backbone=LLaVA-1.5-7B, Vision token count=1442026.0333.9
TokenPackerBackbone=LLaVA-1.5-7B, Vision token count=642026.0333.8
C-AbstractorBackbone=LLaVA-1.5-7B, Vision token count=1442026.0333.4
Pixel-ShuffleBackbone=LLaVA-1.5-7B, Vision token count=1442026.0332.7