MiniMax launches M2.7, a 'self-evolving' AI model.

Chinese AI startup MiniMax has unveiled its new proprietary large language model (LLM), M2.7, which is described as "self-evolving" and capable of handling a significant portion of reinforcement learning research workflows. MiniMax has gained recognition in the global AI market for its frontier-level LLMs, often released with open-source licenses, and its high-quality AI video generation models. The M2.7 model is specifically designed to power AI agents and serve as a backend for third-party tools like Claude Code, Kilo Code, and OpenClaw. This development positions MiniMax as a key player in advancing AI agent capabilities and potentially streamlining complex research processes, further solidifying its reputation for innovation in the competitive AI landscape.