One year after DeepSeek, Chinese AI firms from Alibaba to Moonshot race to release new models

Core Insights - Chinese companies are rapidly advancing their artificial intelligence models in response to competition from U.S. firms like OpenAI, Anthropic, and Google [1] - The introduction of DeepSeek by a Chinese company has raised questions about the effectiveness of U.S. tech restrictions on China [1] Group 1: New AI Model Releases - Moonshot AI launched Kimi K2.5, claiming superior video-generation and agentic capabilities compared to leading U.S. AI models [2] - Alibaba introduced its generative AI model Qwen3-Max-Thinking, which reportedly outperformed U.S. competitors on a benchmark test called "Humanity's Last Exam" [3] - Z.ai released a free version of its GLM 4.7 model, but had to restrict new subscribers for its AI coding tool due to high demand [4] Group 2: Features and Capabilities - Kimi K2.5 aims to create sophisticated agents that can operate autonomously with minimal user interaction [2] - Qwen3-Max-Thinking can generate text, images, or video based on user commands and efficiently selects the best AI tool for various tasks [3] - The rapid rollout of these models indicates a significant push by Chinese companies to enhance their AI capabilities and market presence [1][2][3]