Original sourceThe Decoder
Summary
Baidu released ERNIE 5.5 Flash Base, a 280-billion-parameter open-source foundation model designed for developers to fine-tune after release. It uses a mixture-of-experts architecture activating about 14 billion parameters per input, pretrained on 33.5 trillion tokens, with an original context of 1…
Key points
- Gives teams that need their own post-training another large open foundation model option for fine-tuning.
- Releasing it as a base model returns control over post-training and domain adaptation to developers.
- Teams with fine-tuning capabilities can first validate on small tasks, then assess whether long-context extension fits their use cases.
Editorial note
This page is Code & Chain's editorial summary of public sources. It may be prepared with AI assistance and published through an automated workflow. Refer to the original sources; this content is not investment, legal, or tax advice.