List your company

Vision-Language Modeling on MVision

50.5AccuracyZPPO
15.03624.24333.4542.657Jun 16, 2026

Evaluation results

MethodLinks
ZPPOModel Scale=2B, Prompt Replay Buffer Augmentation=true2026.0650.5
GRPOModel Scale=2B, Prompt Replay Buffer Augmentation=true2026.0643.4
GRPOModel Scale=2B, Prompt Replay Buffer Augmentation=false2026.0643.1
On-DistillModel Scale=2B, Prompt Replay Buffer Augmentation=true2026.0642.7
On-DistillModel Scale=2B, Prompt Replay Buffer Augmentation=false2026.0641.9
Off-DistillModel Scale=2B, Prompt Replay Buffer Augmentation=true2026.0641.8
Off-DistillModel Scale=2B, Prompt Replay Buffer Augmentation=false2026.0640.4
Qwen3.5Model Scale=2B, Prompt Replay Buffer Augmentation=false2026.0638.4
ZPPOModel Scale=0.8B, Prompt Replay Buffer Augmentation=true2026.0629.2
GRPOModel Scale=0.8B, Prompt Replay Buffer Augmentation=true2026.0620.9
GRPOModel Scale=0.8B, Prompt Replay Buffer Augmentation=false2026.0620.1
On-DistillModel Scale=0.8B, Prompt Replay Buffer Augmentation=false2026.0619.4
On-DistillModel Scale=0.8B, Prompt Replay Buffer Augmentation=true2026.0618.7
Off-DistillModel Scale=0.8B, Prompt Replay Buffer Augmentation=true2026.0617.9
Off-DistillModel Scale=0.8B, Prompt Replay Buffer Augmentation=false2026.0617.3
Qwen3.5Model Scale=0.8B, Prompt Replay Buffer Augmentation=false2026.0616.4