Abstract   High-frequency trading (HFT) firms employ sophisticated quantitative models to gain an edge in ultra-fast markets. Among these, high frequency trading Hidden Markov Models (HMMs)  for regime detection and order flow toxicity analysis  are critical for dynamic strategy adaptation. This paper explores cutting-edge techniques used by top HFT firms, focusing on HMM-based regime switching  and advanced toxicity detection methods  beyond conventional microprice forecasti