GRAIN beats multi-agent graph reasoning by 16.45% with one agent and 24% less latency.

LLMs are brittle to renamed nodes and reworded formulations in graph reasoning, and the standard fix is throwing a multi-agent system at the parsing failures. GRAIN is a single RL-trained agent modeling reasoning as semantic parsing plus tool execution, rewarded by a Structure Invariance Reward that validates extracted intermediate graphs against ground-truth topology. It halves the out-of-distribution gap of SFT models from 15.77% to 7.80%. (arXiv 2608.27142) Second result this issue where a trained single agent beat an orchestrated multi-agent baseline.