List your company

One-step next-observation prediction on BabyAI (test)

93Token F1Word2World
38.9252.966781.04May 29, 2026

Evaluation results

MethodLinks
Word2WorldBackbone=Qwen3.5-4B, LLM Usage=SFT, Infer. LLM?=Yes2026.059375
PatchWorld-SimpleBackbone=Qwen3-Coder-480B, LLM Usage=CodeGen, Infer. LLM?=No2026.058558
LLM-DirectBackbone=Mimo-v2.5, LLM Usage=ICL, Infer. LLM?=Yes2026.058154
LLM-DirectBackbone=DeepSeek-V4-Flash, LLM Usage=ICL, Infer. LLM?=Yes2026.058156
PatchWorld-ResidualBackbone=DeepSeek-V4-Flash, LLM Usage=CodeGen, Infer. LLM?=No2026.058056
WorldCoderBackbone=Qwen3-Coder-480B, LLM Usage=CodeGen, Infer. LLM?=No2026.057861
PoE-WorldBackbone=Qwen3-Coder-480B, LLM Usage=CodeGen, Infer. LLM?=No2026.057862
PoE-WorldBackbone=Mimo-v2.5, LLM Usage=CodeGen, Infer. LLM?=No2026.057862
WorldCoderBackbone=Mimo-v2.5, LLM Usage=CodeGen, Infer. LLM?=No2026.057761
PatchWorld-SimpleBackbone=DeepSeek-V4-Flash, LLM Usage=CodeGen, Infer. LLM?=No2026.057751
LLM-DirectBackbone=Qwen3-Coder-480B, LLM Usage=ICL, Infer. LLM?=Yes2026.057344
PoE-WorldBackbone=DeepSeek-V4-Flash, LLM Usage=CodeGen, Infer. LLM?=No2026.057354
WorldCoderBackbone=DeepSeek-V4-Flash, LLM Usage=CodeGen, Infer. LLM?=No2026.057054
PatchWorld-ResidualBackbone=Qwen3-Coder-480B, LLM Usage=CodeGen, Infer. LLM?=No2026.056947
PatchWorld-ResidualBackbone=Mimo-v2.5, LLM Usage=CodeGen, Infer. LLM?=No2026.054928
PatchWorld-SimpleBackbone=Mimo-v2.5, LLM Usage=CodeGen, Infer. LLM?=No2026.054119