Baidu's ERNIE 5.5 Flash Uses 280B-Parameter MoE With 14B Active Per Token

Baidu's ERNIE 5.5 Flash is a 280B-parameter sparse mixture-of-experts model that activates 14B parameters per token, serves over 200 tokens per second, and will ship open weights by November 25.

Read the full story at AlphaSignal ↗

Timeline · 1 report

2026-11-11 05:42 · AlphaSignal
Baidu's ERNIE 5.5 Flash Packs 280B Parameters Into a 14B Compute Path

More stories

Baidu's ERNIE Evolve Automated Research System Beats NanoChat and NanoGPT Records — MarkTechPost
Baidu says AI research agent sped up its next ERNIE model — The Next Web
Baidu's ERNIE Evolve automates AI research, boosting speed sixfold — AlphaSignal
ERNIE-Image 8B: Posters and Chinese Text First, Portraits Second — DEV Community — AI
How to Run Baidu Unlimited-OCR-2 Locally: 3B MIT OCR for 300-Page PDFs — DEV Community — Machine Learning
Anthropic IPO expected to begin trading Nov. 19 at $1.4 trillion valuation — CNBC Technology
Nvidia CEO Jensen Huang faces spy scandal and scrutiny over China AI chip access — New York Times AI
DeepSeek's Harness agent app reaches 1.54 million daily users — Bloomberg AI

Get the daily brief of stories like this at 6:30 every morning →