Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning

Benchmark Model Rank Results
robot-manipulation-on-liberoFast-ThinkAct (3B)–Success Rate: 89.7