Moonshot AI, an artificial intelligence firm supported by Alibaba, has launched Kimi K2.5, the latest upgrade of its large language model (LLM), as domestic competition in China’s AI market escalates. The new release arrives amid a surge of technology rollouts from leading Chinese companies striving to match the capabilities of Western AI developers.
Moonshot’s Kimi K2.5 model introduces the capability to handle text, images, and video inputs within a single prompt, reflecting a broader shift toward multimodal AI systems led by international players such as OpenAI and Google. The company has indicated that the latest version includes enhancements in coding, enabling the generation of functional and visually detailed user interfaces from natural language commands.
Performance assessments reveal K2.5 surpasses open-source equivalents on several industry benchmarks, while closing in on proprietary leaders in coding tasks. Alongside this launch, Moonshot is introducing an automated coding tool designed to compete with Anthropic’s Claude Code offering.
The company notes that Kimi K2.5 is compatible with OpenAI’s API protocols, aimed at lowering transition barriers for developers.
The announcement comes as multiple Chinese AI firms accelerate releases, anticipating forthcoming advancements from market peers such as DeepSeek, which has signaled a significant update is imminent and has been active in publishing technical research.
Investor interest in Chinese AI model development remains high. Last month, Moonshot secured $500 million in funding from investors including Alibaba and IDG Capital, with its valuation reaching $4.3 billion.
With the debut of Kimi K2.5, Moonshot and other major Chinese AI developers, backed by influential technology investors, continue to push for technological parity with Western AI models and expand their user base in a highly competitive global landscape.
