🤖 Chinese MiniMax claims its AI outperforms DeepSeek

Shanghai-based AI startup MiniMax has released its new language model, MiniMax-M1, claiming it outperforms its rivals — including Chinese market leader DeepSeek.

📋 Key features

⚫️ M1 supports up to 1 million tokens of context — 8x more than DeepSeek R1 ⚫️ In some tasks, M1 uses 70% less computational resources ⚫️ Trained using reinforcement learning on 512 Nvidia H800 GPUs (rental cost: $534,700)

📌 Why it matters

⚫️ MiniMax is one of the six “Little Dragons” of China’s AI market ⚫️ Backed by Tencent and Alibaba ⚫️ While most competitors have scaled back core research under DeepSeek’s dominance, MiniMax is taking a different path

The company is also developing a video generator and AI companion apps — with more announcements expected in the coming days.