China's open-weight Kimi model stuns AI world with frontier-level results

The Chinese startup Moonshot AI has released a new open-weight version of its Kimi large language model that demonstrates performance metrics comparable to top-tier international models. By providing high-capability weights for public use, the company has notably narrowed the performance gap between domestic Chinese AI tools and global frontier models. This development increases competitive pressure in the open-model ecosystem and highlights the rapid technical progress being made by Chinese firms despite ongoing restrictions on advanced hardware.