Baidu releases ERNIE 5.5 Flash Base with 280 billion parameters as an open-source foundation model

Baidu has released ERNIE 5.5 Flash Base, a 280-billion-parameter model utilizing a mixture-of-experts architecture. The model is pre-trained on 33.5 trillion tokens and is being distributed in an un-tuned state. By providing the base model without specific task-oriented training, Baidu allows developers to customize the architecture for their own applications. This move provides researchers and engineers with a large-scale foundation model that can be adapted for a wide variety of specialized use cases.