Chinese open-source AI models gain users without market disruption

Chinese companies released open-source AI models last month that rank among the highest-performing freely available systems. ai placed fifth on OpenRouter usage rankings and first among open-source options according to Artificial Analysis benchmarks.

The releases produced no measurable drop in U.S. technology company valuations. Earlier models from the same firms had triggered immediate legislative proposals and share-price losses exceeding one trillion dollars.

Different release strategies Chinese firms distribute models for local download at no cost. U.S. companies typically host models in the cloud and charge for access while restricting internal details. Open-source models from China trail the most expensive closed models on some tests but match or exceed them on engineering benchmarks.

A UK software developer reported that a local Chinese model completed the same task in 30 percent more time than a paid cloud model but incurred no usage fees.

Usage and performance data DeepSeek's latest model leads OpenRouter usage with more than twice the traffic of the next model. Seven of the ten most-used models on the platform originate from Chinese companies. Artificial Analysis ranked GLM-5.2 as the top open-source model but noted it scored slightly below GPT-5.5 on intelligence tests.

The firm stated that no industry-wide standard exists for measuring AI performance.

Policy and market context U.S. export controls on advanced chips to China remain in place. Some analysts noted that restricting access to U.S. models may accelerate development of independent Chinese systems. Corporate adoption of open-source models faces internal compliance and risk-review hurdles.

Large established vendors continue to supply enterprise software under existing contracts that address regulatory and security requirements.