List your company

Interactive Task Solving on ALFWorld Large context window

100Pick Success RateEmbedding
78.36883.98489.695.216Jun 10, 2026

Evaluation results

MethodLinks
EmbeddingBackbone=Claude Sonnet 4.5, Memory Type=Internal Static Memory, Context Input Token Limit=2200 tokens2026.061003.250077.835.342.5
ReSumBackbone=Claude Sonnet 4.5, Memory Type=Internal Dynamic Memory, Context Input Token Limit=2200 tokens2026.0610067.778.333.372.276.571.6
Slide WindowBackbone=Claude Sonnet 4.5, Memory Type=Internal Static Memory, Context Input Token Limit=2200 tokens2026.0695.861.382.614.383.388.270.1
FoldBackbone=Claude Sonnet 4.5, Memory Type=Internal Dynamic Memory, Context Input Token Limit=2200 tokens2026.0695.835.587077.864.759
HIAGENTBackbone=Claude Sonnet 4.5, Memory Type=Internal Dynamic Memory, Context Input Token Limit=2200 tokens2026.0695.854.878.314.377.870.664.9
AconBackbone=Claude Sonnet 4.5, Memory Type=Internal Dynamic Memory, Context Input Token Limit=2200 tokens2026.0695.86.547.8077.841.242.5
TruncationBackbone=Claude Sonnet 4.5, Memory Type=Internal Static Memory, Context Input Token Limit=2200 tokens2026.0691.7047.84.855.611.834.3
HORMABackbone=Claude Sonnet 4.5, Memory Type=External Dynamic Memory, Context Input Token Limit=2200 tokens2026.0679.283.960.995.266.747.173.9