List your company

Real-world perception-centric reasoning on RealWorldQA (test)

77.04AccuracyGLM-9B-DeltaThinker
63.852867.276470.774.1236May 15, 2026

Evaluation results

MethodLinks
GLM-9B-DeltaThinkerAlgorithm=DeltaPrompts2026.0577.04
Qwen-8B-DeltaThinkerAlgorithm=DeltaPrompts2026.0575.82
Qwen3-VL-8B-Thinking2026.0573.07
GLM-4.1V-9B-Thinking2026.0572.19
Bee-8B-RL2026.0572.03
Vision-R1-7B2026.0567.58
ARES-RL-7B2026.0566.67
REVisual-R12026.0564.36