List your company

Interactive Agent on SciWorld (Pass@1, Pass@128)

29.45Pass@1ACE
6.10212.163518.22524.2865Mar 17, 2026

Evaluation results

MethodLinks
ACEBackbone=Qwen2.5-7B2026.0329.4559.67
LEAFEBackbone=Qwen2.5-7B2026.0327.8862
GRPO-RLVRBackbone=Qwen2.5-7B2026.0327.1757.33
EARLY-EXPBackbone=Qwen2.5-7B2026.0326.1754.67
ACEBackbone=Llama3.1-8B2026.0325.2857.33
GRPO-RLVRBackbone=Llama3.1-8B2026.0324.2556
EARLY-EXPBackbone=Llama3.1-8B2026.0324.0456
LEAFEBackbone=Llama3.1-8B2026.0322.759.33
BASE (NO FT)Backbone=Llama3.1-8B2026.037.1748.67
BASE (NO FT)Backbone=Qwen2.5-7B2026.03747.33