Chinese technology companies are speeding up the release of new artificial intelligence models as competition with U.S. firms intensifies. The push comes just over a year after DeepSeek drew global attention with a low-cost AI chatbot.
Beijing-based startup Moonshot AI this week unveiled Kimi K2.5, an upgraded model with video generation and agent-style task abilities. The company said the model performs better than leading systems from OpenAI, Google, and Anthropic. The release followed Moonshot’s K2 model by only three months.
Alibaba also announced a new generative AI model called Qwen3-Max-Thinking. The company said it performed strongly on a major benchmark test and can handle text, image, and video tasks. Alibaba added that the model works efficiently by choosing the best tools automatically.
Earlier in January, Z.ai launched a free version of its GLM 4.7 model. Strong demand forced the company to limit new sign-ups for its AI coding tool. Baidu also released an updated model, Ernie 5.0, which helped lift its stock to a near three-year high.
Experts say China’s AI models are now only months behind those developed in the United States. Many Chinese companies focus on rapid updates and wider access rather than only building the most advanced systems.
Chinese firms are also competing by offering open-source or low-cost models. This approach helps attract users in emerging markets. Microsoft recently noted that DeepSeek usage in parts of Africa is far higher than in other regions.
Tech giants like Tencent, Alibaba, and ByteDance are tying AI tools into existing apps and services. Promotions and in-app rewards aim to boost user traffic and long-term adoption.
Analysts say the strategy is clear. Chinese companies want more users and deeper integration into daily digital life, even if it means prioritizing reach over cutting-edge performance.
