List your company

E-commerce product search and purchase on WebShop

81.3Strict SuccessSIRI
0.80421.70242.663.498May 14, 2026May 21, 2026May 29, 2026Jun 6, 2026Jun 14, 2026Jun 22, 2026Jun 30, 2026

Evaluation results

MethodLinks
SIRICategory=Memory-Augmented RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0681.3—89.9
TRIAGEModel=Qwen2.5-7B-Instruct, Judge=Qwen3-8B-thinking, Evidence=true2026.0677.2——
TRIAGEModel=Qwen2.5-7B-Instruct, Judge=Qwen3-8B-thinking, Evidence=false2026.0673.5——
GiGPOCategory=RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0672.8—84.4
SkillRLCategory=Memory-Augmented RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0672.7—85.2
GRPOModel=Qwen2.5-7B-Instruct2026.0670.1——
PPOCategory=RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0668.7—81.4
GRPOCategory=RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0666.1—79.3
RLOOCategory=RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0665.7—80.3
TRIAGEModel=Qwen2.5-7B-Instruct, Judge=Qwen3-8B no-think2026.0665.4——
TRIAGEModel=Qwen3-1.7B-Instruct, Judge=Qwen3-8B-thinking, Evidence=true2026.0655.9——
TRIAGEModel=Qwen3-1.7B-Instruct, Judge=Qwen3-8B-thinking, Evidence=false2026.0649.6——
SimpleMem+GRPOCategory=Memory-Augmented RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0646.9—67.8
Mem0+GRPOCategory=Memory-Augmented RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0637.5—58.1
GRPOModel=Qwen3-1.7B-Instruct2026.0637.5——
ActFocusBackbone=3B, Evaluation Step=1002026.0536.799.2—
TRIAGEModel=Qwen3-1.7B-Instruct, Judge=Qwen3-8B no-think2026.0635.1——
ReflexionCategory=Prompt-based Agentic or Memory-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0628.8—58.1
ActFocusBackbone=1.5B, Evaluation Step=1002026.0526.698.8—
UniformBackbone=3B, Evaluation Step=1002026.0522.796.9—
ReActCategory=Prompt-based Agentic or Memory-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0619.5—46.2
EvolverCategory=Memory-Augmented RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.0617.6—42.5
Gemini-3-FlashCategory=Closed-source LLMs2026.0616.5—14.1
GLM-5.1Category=Closed-source LLMs2026.0612.5—15.6
UniformBackbone=1.5B, Evaluation Step=1002026.059.894.9—
MemRLCategory=Memory-Augmented RL-based Methods, Backbone=Qwen2.5-7B-Instruct2026.069.2—29.5
OriginBackbone=Qwen2.5-7B-Instruct2026.063.9—16.6