SAM: State-Adaptive Memory for Long-Horizon Reasoning Agent
Open paper
Benchmark
Model
Rank
Results
deep-research-agents-on-browsecomp-zh
GLM-4.7 with SAM
#4
Accuracy: 64.2
deep-research-agents-on-browsecomp
GLM-4.7 + SAM
#39
Accuracy: 56.5
Rank counts only results with a code link.