Unbiased AI-powered news
Moonshot's Kimi K3 model topped Arena's coding capability ranking. The Beijing startup priced access at half the cost of OpenAI's top model.
pymnts.comA Beijing-based startup released a new artificial intelligence model that ranked first in coding capability tests, according to Arena, a platform that evaluates AI systems. The Kimi K3 model from Moonshot outperformed earlier versions of Anthropic's Claude and OpenAI's ChatGPT on the platform's front-end coding benchmark.
Arena co-founder and CEO Anastasios Angelopoulos called the release the single biggest of the year.
America research analysts reported that K3 costs half as much to use as OpenAI's GPT-5.6 Sol model. Moonshot has not disclosed the hardware used to train K3, though the company partners with Huawei. The release occurred one day before Chinese President Xi Jinping's address at the World Artificial Intelligence Conference in Shanghai. The conference continues through Monday.
K3 follows a May release of GLM-5.2 from Chinese startup Zhipu.
San Francisco-based Anysphere has incorporated Moonshot's earlier K2.5 model into its Cursor coding tool. SpaceX plans to purchase Cursor for $60 billion later this year. Moonshot co-founder Yang Zhilin earned his doctorate at Carnegie Mellon University in 2019.
These outlets didn't split into competing frames — coverage was uniform.
Chinese President Xi Jinping addressed the World Artificial Intelligence Conference in Shanghai on July 17, 2026. He presented China's lower-cost AI approach as an alternative to U.S. models and urged international cooperation.
Al JazeeraChinese President Xi Jinping urged international cooperation on artificial intelligence and warned against any single country dominating the field during a keynote address at the World Artificial Intelligence Conference in Shanghai. He also announced plans to work with partners i…
zerohedge.comMeta is negotiating a multi-year agreement to supply computing capacity from its data centers to Anthropic. The arrangement could reach $10 billion over two years and would mark a new revenue stream for Meta while addressing shortages of AI infrastructure.