MiniMax M3 AI: Outperforms GPT-5.5, Cuts Costs

MiniMax, a Chinese AI start-up, has launched its new M3 AI model, which can outperform OpenAI's GPT-5.5. The company introduced M3 as part of its expansion into coding agents and automated workflows. Here's the thing: M3's innovative architecture significantly cuts down inference costs and improves response speeds by reducing computational requirements by up to 95%. What's interesting is that M3 can process one million tokens of data at once, which is five times more than its predecessor. This allows it to handle long and complex programming tasks. In a major coding benchmark, SWE-Bench Pro, M3 surpassed rival models like OpenAI's GPT-5.5 and Google's Gemini 3.1 Pro. This shows its high capability for software engineering and automated tasks. The launch of M3 is MiniMax's first major product release since it began preparing for an IPO on Shanghai's Star Market. This development highlights the increasing competition in the global AI landscape.