List your company

Tracking Shuffled Objects on Tracking Shuffled Objects 5 objects (test)

94.1Accuracy (TSO 5-obj)TextReg
33.2649.05564.8580.645May 20, 2026

Evaluation results

MethodLinks
TextRegTest Engine=Phi-3.5-Mini-Instruct2026.0594.1
CoTTest Engine=Phi-3.5-Mini-Instruct2026.0589.6
REVOLVETest Engine=Phi-3.5-Mini-Instruct2026.0586.8
TextGradTest Engine=Phi-3.5-Mini-Instruct2026.0583.6
TextRegTest Engine=Llama-3.1-8B-Instruct2026.0579.7
REVOLVETest Engine=Llama-3.1-8B-Instruct2026.0576
TextGradTest Engine=Llama-3.1-8B-Instruct2026.0569.7
CoTTest Engine=Llama-3.1-8B-Instruct2026.0558.8
TextRegTest Engine=Llama-3-8B-Instruct2026.0554.3
REVOLVETest Engine=Llama-3-8B-Instruct2026.0546.1
TextGradTest Engine=Llama-3-8B-Instruct2026.0545.9
TextRegTest Engine=Qwen2-7B-Instruct2026.0545.4
CoTTest Engine=Qwen2-7B-Instruct2026.0542
REVOLVETest Engine=Qwen2-7B-Instruct2026.0540.3
TextGradTest Engine=Qwen2-7B-Instruct2026.0538.2
CoTTest Engine=Llama-3-8B-Instruct2026.0535.6