List your company
ResearchFormerly WizWand

Science Experimentation Reasoning on ScienceWorld (seen)

89.7Success RateSKILLPYRAMID
53.40462.82772.2581.673Jun 2, 2026Jun 7, 2026Jun 13, 2026Jun 19, 2026Jun 25, 2026Jul 1, 2026Jul 7, 2026
Updated 3mo ago

Evaluation results

MethodLinks
SKILLPYRAMIDModel=Gemini 2.5 Pro2026.0689.79
SKILLPYRAMIDModel=GPT-4.12026.0686.615.2
SKILLPYRAMIDModel=Qwen3-235B2026.0685.614
ReAct + SkillsModel=GPT-4.12026.0678.617.6
SkillRerankerBackbone Model=DeepSeek-v4-Flash2026.0778.4615.77
ReAct + SkillsModel=Gemini 2.5 Pro2026.0677.316.7
SKILLPYRAMIDModel=DeepSeek-V3.22026.0676.316.7
SkillRouterBackbone Model=DeepSeek-v4-Flash2026.0775.2816.61
ExpeLModel=DeepSeek-V3.22026.0674.916
ReAct + SkillsModel=Qwen3-235B2026.0674.215.5
SkillRerankerBackbone Model=GPT-5.4-Mini2026.0773.6114.48
Graph of SkillsBackbone Model=DeepSeek-v4-Flash2026.0773.3218.12
LLM-as-selectorBackbone Model=DeepSeek-v4-Flash2026.0773.0617.11
SkillRouterBackbone Model=GPT-5.4-Mini2026.0773.0614.71
SkillRouterBackbone Model=Qwen3.6-27B2026.0772.8512.15
ExpeLModel=Gemini 2.5 Pro2026.0672.815
ReflexionModel=DeepSeek-V3.22026.0672.416.3
SkillRerankerBackbone Model=Qwen3.6-27B2026.0772.2611.26
ExpeLModel=GPT-4.12026.0672.114.4
ReAct + SkillsModel=DeepSeek-V3.22026.0671.615.9
ReActModel=DeepSeek-V3.22026.0669.917.6
ExpeLModel=Qwen3-235B2026.0669.415.3
LLM-as-selectorBackbone Model=Qwen3.6-27B2026.0768.4511.95
Graph of SkillsBackbone Model=Qwen3.6-27B2026.0767.412.42
ReflexionModel=GPT-4.12026.0666.817
ReflexionModel=Gemini 2.5 Pro2026.0666.816.9
LLM-as-selectorBackbone Model=GPT-5.4-Mini2026.0766.5415.59
Graph of SkillsBackbone Model=GPT-5.4-Mini2026.0763.6116.22
ReflexionModel=Qwen3-235B2026.0663.617.7
ReActModel=Gemini 2.5 Pro2026.0658.218.4
ReActModel=GPT-4.12026.0657.419.1
ReActModel=Qwen3-235B2026.0654.819.6