174 likesNew · 15d old
Baidu drops 280B MoE LLM with 14B active tokens
What is ERNIE-5.5-Flash-Base?
ERNIE-5.5-Flash-Base is a 280-billion-parameter mixture-of-experts language model that only activates 14 billion parameters for each token. It can read up to a million tokens at once and was trained on 33.5 trillion tokens for code, reasoning and multilingual tasks.
Who it is for
Researchers and developers building long-form chat, code assistants or domain-specific models can fine-tune it to get strong performance without the full compute cost.