Cursor has released Composer 2.5, an upgraded version of its AI coding assistant built on Kimi K2.5. The update promises improved performance on long-running tasks and better adherence to complex instructions.
Cursor's latest Composer iteration represents a substantial improvement over its predecessor, Composer 2, according to the company. The new version is engineered to handle sustained work on extended coding projects more effectively while maintaining consistency when following intricate user directives.
The foundation for Composer 2.5 is Kimi K2.5, indicating Cursor's integration of the latest language model capabilities. This pairing aims to enhance the assistant's ability to manage multi-step development tasks that require sustained focus and precision.
Key improvements in Composer 2.5 address two critical challenges in AI-assisted coding: maintaining context and accuracy over longer work sessions, and reliably interpreting complex, multi-layered instructions. Developers working on substantial features or refactoring projects may find these enhancements particularly relevant.
The release is available now within the Cursor platform. Users currently running earlier versions can access the update through standard channels. Cursor has not announced pricing changes or additional feature gates tied to this release.
Composer 2.5 enters a competitive landscape where multiple AI coding tools compete on performance metrics, including instruction-following accuracy and context retention. The update reflects ongoing iteration in this space, where language model improvements are frequently translated into more capable coding assistants.
Z.ai released GLM-5.3's weights on Hugging Face under a new license that requires large companies to undergo security review before hosting the model. The change marks a departure from the standard MIT license.
Anthropic has introduced the Model Hardware Standard (MHS), a unified interface enabling AI agents to operate robotic arms, lab instruments, and other physical devices. Early testing shows integration time has dropped from weeks to hours.
Open-weight AI companies—those releasing freely available models—are attracting major acquisition interest from tech giants. The trend reflects growing capital investment in the business model of distributing AI models at no cost.
Google Deepmind has upgraded its Co-Scientist AI system to autonomously plan experiments, operate lab equipment, and publish scientific papers. The Gemini-based multi-agent platform demonstrated experimentally validated results across materials science, chemistry, and medical AI development.