Chinese AI model "MiniMax M3" debuts as a high-performance open model competing with GPT-5.5 and Claude Opus 4.7

Shanghai-based MiniMax unveiled MiniMax M3, a high-performance AI model built for coding, agent tasks, and multimodal use.

The company says M3 can handle up to 1 million tokens of context and accepts both image and video inputs.

On some benchmarks, M3 reportedly outperformed GPT-5.5, Claude Opus 4.7, and Gemini 3.1 Pro.

MiniMax said it will release the model as an open model within 10 days, highlighting a lower-cost alternative to top closed models.