Yandex Refutes Rumors That Its Neural Networks Are Based on China's Qwen
Яндекс
Yandex has published an official statement refuting rumors that its neural networks (YandexGPT, YandexART, and others) are based on the open-source Chinese model Qwen. The company claims that it uses its own architectural solutions and trains models on its own data.
Yandex has officially denied information that its neural networks (including YandexGPT and YandexART) are based on open-source Chinese Qwen models. Earlier, claims appeared online about an alleged dependency, but company representatives stated that this is not true. Yandex emphasized that all key technologies are developed in-house, using proprietary architectural solutions and training on their own data. The company also noted that, like many other developers, it may use open models as one of the stages, but the final model is entirely original.
Source: Яндекс ИИ (GNews) —
original