Ring-2.6-1T Officially Open-Sourced: Ant Ling Releases Trillion-Scale Thinking Model
Original title:Ring-2.6-1T 正式开源:为真实复杂任务打造的万亿级思考模型
AI summary
Ant Ling open-sourced Ring-2.6-1T, a trillion-scale thinking model focused on strengthening agent execution capabilities, achieving open-source SOTA on PinchBench and ClawEval. The model supports two reasoning intensity levels, high and xhigh, and adopts an asynchronous Async RL architecture combined with an upgraded Popsicle algorithm to improve training efficiency. Weights have been open-sourced on Hugging Face and ModelScope, and the Chat experience is now available.
Key Takeaways
- Ant Ling open-sourced Ring-2.
- 6-1T,a trillion-scale thinking model focused on strengthening agent execution…
- The model supports two reasoning intensity levels,high and xhigh,and adopts an asynchronous Async…
- Weights have been open-sourced on Hugging Face and ModelScope,and the Chat experience is now…
Quick Facts
- Published
- Updated
- Source
- 公众号:蚂蚁百灵(Ling)
- Category
- Model Releases
- AI score
- 80
通过可调节推理强度,模型在 Agent 高频任务中减少不必要计算,复杂场景又能释放上限,为万亿参数模型的实际部署提供了更精细的资源控制方式。
Related entities
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: Ring-2.6-1T Officially Open-Sourced: Ant Ling Releases Trillion-Scale Thinking Model. May 15, 2026. https://new.nat6.net/en/news/cmsdu2qga010krofzw3q4wtj2/