MiniMax releases M3 with MSA architecture, 1M-token context, native multimodality, and frontier-level coding and agentic performance.

Shanghai-based company says M3 can process data five times faster than its predecessor, while also slashing inference costs.

Chinese start-up MiniMax has launched M3, an AI model with a redesigned architecture that reduces computational needs by up to 95%, enhancing efficiency and response speeds.