baidu/ERNIE-5.5-Flash-Base: Sparse Mixture-of-Experts Language Model Release

Baidu has released ERNIE-5.5-Flash-Base, an efficient sparse Mixture-of-Experts language model featuring 280B total parameters and 14B activated parameters per token. Pre-trained on 33.5T tokens with a context length supporting up to 1M tokens for fine-tuning and inference, the model utilizes hybrid attention and multi-router latent MoE designs to deliver strong performance across reasoning, math, and code benchmarks. Open article →