According to a research report published by Jefferies, Minimax recently officially released the M3.1 Flash model, which has significantly improved capabilities in various fields such as agentic tasks, programming, computer usage, and network security, and has strong cost efficiency. The M3.1 Flash model's inference speed is increased by about 3 times, and the cost is reduced by 70%. It is expected to redefine the capabilities and cost efficiency of AI agents. The bank pointed out that Minimax emphasizes that the Flash version focuses on cost efficiency and performance balance, and the Pro version will be launched later, focusing on smart limits. It is expected that the sparse attention architecture will drive better reasoning efficiency. The bank maintains the Minimax “buy” rating and target price of HK$533.

According to a research report published by Jefferies, Minimax recently officially released the M3.1 Flash model, which has significantly improved capabilities in various fields such as agentic tasks, programming, computer usage, and network security, and has strong cost efficiency. The M3.1 Flash model's inference speed is increased by about 3 times, and the cost is reduced by 70%. It is expected to redefine the capabilities and cost efficiency of AI agents. The bank pointed out that Minimax emphasizes that the Flash version focuses on cost efficiency and performance balance, and the Pro version will be launched later, focusing on smart limits. It is expected that the sparse attention architecture will drive better reasoning efficiency. The bank maintains the Minimax “buy” rating and target price of HK$533.