Ant Group's inclusionAI open-sources Ling-3.0 series language foundation models
Original title:inclusionAI/Ling-3.0-tiny-base
AI summary
Ant Group's inclusionAI open-sourced the Ling-3.0 series language foundation models, including tiny and flash sizes and checkpoints from pretraining, mid-training and merging (WSM) stages. The series uses a hybrid linear attention and sparse MoE architecture with 7.9B total parameters, activating only 1.3B parameters per token (8 of 128 routed experts).
Key Takeaways
- Ant Group's inclusionAI open-sourced the Ling-3.
- 0 series language foundation models,including tiny and flash sizes and checkpoints from…
- The series uses a hybrid linear attention and sparse MoE architecture with 7.
- 9B total parameters,activating only 1.
- 3B parameters per token (8 of 128 routed experts).
Quick Facts
- Published
- Updated
- Category
- Model Releases
- AI score
- 65
用加权检查点合并替代学习率衰减,让基座模型更适合持续预训练,也为验证不同衰减曲线省去重复实验。
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Ant Group's inclusionAI open-sources Ling-3.0 series language foundation models. August 11, 2026. https://new.nat6.net/en/news/cmt0aa8b416ywrodpq8czlstx/