FrontierThe story, in brief
Baidu's Ernie 5.1 cuts 94 percent of pre-training costs while competing with top models
94%. That's how much Baidu just slashed pre-training costs with Ernie 5.1—and it's still ranking 4th globally.
The KeyNews take
Why it matters
Baidu's efficiency breakthrough challenges the compute-scaling paradigm that's dominated AI development. A 'Once-For-All' training approach that extracts multiple sub-models in a single run could reshape how labs think about training ROI and competitive positioning.
The key facts
6 to knowErnie 5.1 uses 1/3 the parameters of predecessor
Pre-training cost: 6% of comparable models (94% reduction)
Training approach: 'Once-For-All' methodology extracting sub-models from single run
Search Arena leaderboard rank: 4th globally
Competitors ranked higher: Claude Opus (2 variants), GPT-5.5 Search
Published: May 11, 2026
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Baidu's Ernie 5.1 uses just a third of its predecessor's parameters and reportedly cost only six percent of what comparable models require to pre-train. That's possible thanks to a "Once-For-All" approach that extracts smaller sub-models from a single training run. On the Search Arena leaderboard,…