Chinese companies are releasing new AI models and striving to gain an edge in the AI application race as the country's biggest holiday of the year nears.

On the Humanity’s Last Exam (HLE) benchmark, Kimi K2.5 scored 50.2% (with tools), surpassing OpenAI’s GPT-5.2 (xhigh) and Claude Opus 4.5. It also achieved 76.8% on SWE-bench Verified, cementing its status as a top-tier coding model, though GPT-5.2 and Opus 4.5 overtake it here at 80 and 80.9, respectively.

OpenAI, Google, and Moonshot AI are ushering in agentic AI systems that investigate, coordinate, and verify tasks beyond simple prompts.

The company said that the model was trained on 15 trillion mixed visual and text tokens.

Kimi has a standard mode and a Thinking mode that offers higher output quality. Additionally, a capability called K2.5 Agent Swarm enables the LLM to split complex tasks into simpler sub-steps and assign each sub-step to a separate AI agent. A built-in orchestration engine can create and manage up to 100 agents per prompt.

Alibaba Group Holding Ltd.-backed Moonshot AI released an upgrade of its flagship model, heating up a domestic arms race ahead of an expected rollout by Chinese sensation DeepSeek.

Moonshot AI, maker of Kimi, is valued near $4.8B after a $500M jump in weeks, as China’s AI market reprices amid surging interest in local chatbots.

A company that makes equipment critical to the artificial intelligence boom is investing tens of millions to open a new headquarters and production facility in Lewisville. In this story, read about Moonshot's meteoric growth, including a revenue surge of more than 3,800%, and hear from CEO Ethan Ellenberg about its future plans.

Is your CEO being overpaid for underperformance? New research suggests it's systemic, and jumbo pay packages make it worse.

In today’s CEO Daily: Fortune Editor-in-Chief Alyson Shontell interviews Pfizer CEO Albert Bourla. The big story: Target’s incoming CEO speaks out about the immigration crackdown in the retailer’s home state.