List your company

Perception on HRBen 8K (Pass@4)

90Pass@4AXPO
71.876.52581.2585.975May 27, 2026

Evaluation results

MethodLinks
AXPOModel Size=4B, Backbone=Qwen3-VL-Thinking2026.0590
BaseModel Size=32B, Backbone=Qwen3-VL-Thinking2026.0590
AXPOModel Size=8B, Backbone=Qwen3-VL-Thinking2026.0589
SFTModel Size=2B, Backbone=Qwen3-VL-Thinking2026.0588
AXPOModel Size=2B, Backbone=Qwen3-VL-Thinking2026.0588
SFTModel Size=4B, Backbone=Qwen3-VL-Thinking2026.0588
SFTModel Size=8B, Backbone=Qwen3-VL-Thinking2026.0587
SFT + GRPOModel Size=2B, Backbone=Qwen3-VL-Thinking2026.0585.5
GRPOModel Size=4B, Backbone=Qwen3-VL-Thinking2026.0585.5
SFT + GRPOModel Size=4B, Backbone=Qwen3-VL-Thinking2026.0584.5
SFT + GRPOModel Size=8B, Backbone=Qwen3-VL-Thinking2026.0584.5
BaseModel Size=8B, Backbone=Qwen3-VL-Thinking2026.0582.5
GRPOModel Size=8B, Backbone=Qwen3-VL-Thinking2026.0582
BaseModel Size=4B, Backbone=Qwen3-VL-Thinking2026.0581
BaseModel Size=2B, Backbone=Qwen3-VL-Thinking2026.0577
GRPOModel Size=2B, Backbone=Qwen3-VL-Thinking2026.0572.5