Kimi K3 Isn't the Cheapest AI Model, Test Shows

A switching-cost stress test published by tech publication TechLoy shows that Moonshot AI's Kimi K3 is not the cheapest AI model available.

TechLoy describes the test as modeling a scenario in which a business loses access to its current model and moves the same monthly token workload to another provider, then measures how its API bill would change. The conclusion, as the headline puts it: Kimi K3 is not the cheapest option.

The test arrives amid intensifying debate over US restrictions on Chinese AI models. TechLoy notes the US has not announced a blanket ban on Chinese models, but officials have considered sanctions and other restrictions over concerns about intellectual property, access to advanced chips, and the security of powerful open-weight models — with the immediate dispute centred on Moonshot AI, the company behind Kimi K3.

Chinese models have also become increasingly central to developer ecosystems. According to data cited by Yahoo Finance, Chinese models account for about 63% of token usage by US companies on OpenRouter, a platform that lets developers access multiple AI models through one API.

In that context, the finding that Kimi K3 is not the cheapest option carries extra weight: pricing competition among Chinese open-source models is far from settled, and no single vendor can claim a monopoly on cost advantage.

Watch for how US policy toward Chinese AI developers evolves, how Chinese model vendors adjust pricing, and whether the open-source model market enters a new round of price competition.