Moonshot AI’s Kimi K3 Model Challenges Top US AI Systems in New Rivalry

Moonshot’s Kimi K3 Challenges Anthropic and OpenAI

Moonshot’s Kimi K3, run by a Pink Floyd-loving entrepreneur who earned his doctorate in Pittsburgh, appears to be catching up to the best versions of Anthropic’s Claude and OpenAI’s ChatGPT. The release occurred shortly before Chinese President Xi Jinping’s opening address to the nation’s annual World Artificial Intelligence Conference in Shanghai, where Xi stated, The development of artificial intelligence should not be a solo performance by any single country but rather a symphony of global cooperation.

Anastasios Angelopoulos, the co-founder and CEO of Arena, a platform for evaluating AI systems, described the release as the single biggest release of the year, and a moment when open-source Chinese models are surpassing closed U.S. models. Angelopoulos noted that Kimi K3 topped the charts in Arena’s ranking of front-end coding capability. He added, More results are rolling in that are likely to continue to show it is at the top of the pack.

The model is a massive 2.8 trillion parameters, a whopping 20 times the size of Poolside’s new release. While it carries significant momentum, tech analyst Patrick Moorhead noted on social media that the response to K3 is an overreaction shockingly similar to the release of DeepSeek’s model in early 2025. He stated it could be good for parts of the broader AI industry but poses a revenue challenge to Anthropic and OpenAI.

Poolside Deploys Industrial Model-Building Process for Laguna

In response to the dominance of Chinese labs, U.S. startups are recalibrating their strategies. Poolside, co-founded by former GitHub CTO Jason Warner, recently debuted its Laguna model. Warner, who serves as co-CEO, tells Forbes, As an American company building for the West, we’ll be the most capable open model in the West. Globally, in this weight class of the 118 billion parameter model, we are the leader. Warner claims the startup spent 18 months building the infrastructure for a model building factory that could pump out Laguna and continue with new iterations every five weeks. The classic notion of model building is an artisanal process…We've built an industrial model-building process, he says.

Mira Murati’s Thinking Machines Lab Introduces Inkling

Simultaneously, Thinking Machines Lab, the startup founded by former OpenAI CTO Mira Murati, has introduced Inkling, a Mixture-of-Experts Transformer with 975 billion total parameters, 41 billion of which are active at any given time. Positioned as a flexible base for fine-tuning, Inkling natively handles text, images, and audio, and supports a context window of up to one million tokens. The company is positioning Inkling as a flexible base model for customization, though the announcement states, Inkling is not the strongest overall model available today.

Data from Artificial Analysis provides a granular look at how these U.S. models compare to their Chinese rivals. Inkling debuts with a score of 41 on the Artificial Analysis Intelligence Index, making it the leading open-weights model from a U.S. lab. It ranks three points above the previous leader, Nemotron 3 Ultra at 38, and well ahead of Gemma 4 31B at 29 and gpt-oss-120b at 24. On GDPval-AA v2, an agent-based benchmark, Inkling reaches an Elo rating of 1,238, beating Kimi K2.6 at 1,190 and DeepSeek v4 Flash max at 1,189. Inkling also scores 24 percent on the Tau-3 banking benchmark, ahead of Kimi K2.6 at 21 percent and DeepSeek v4 Flash max at 23 percent.

Thinking Machines says it pre-trained Inkling on 45 trillion tokens of public and synthetic text, images, audio recordings, and videos. The company used the Chinese AI model Kimi K2.5, among other methods, to generate synthetic data. Kimi K2.5 also served as the basis for Cursor’s coding model.

The race is as much about infrastructure as it is about algorithms. During the World Artificial Intelligence Conference, tech giant Huawei has been showcasing a new AI computing system called the Atlas 950 SuperPoD, a signal that China increasingly is amassing the domestic hardware it needs despite U.S. restrictions on imports from chipmakers like Nvidia. K3 follows another major AI model release last month from the Chinese startup Zhipu, or Z.ai. Its new flagship GLM-5.2 model is already widely used by software developers who say it can perform work almost as well as top U.S. models at a lower price.

Sources: Forbes, The Decoder.