Model Releases

Ant Ling Releases Ling & Ring 2.6 Technical Report

Original title:百灵 Ling & Ring 2.6 技术报告发布:面向真实 Agent 工作流的高效万亿级模型

AI summary

Ant Ling released the Ling & Ring 2.6 technical report, systematically disclosing the architecture, pre-training, post-training and agent reinforcement learning details of Ling-2.6-flash, Ling-2.6-1T and Ring-2.6-1T. The three models adopt a Hybrid Linear Attention architecture that combines Lightning Attention and MLA at a 7:1 ratio. Ling-2.6-flash achieves a decoding speed of 340 tokens/s on 4×H20 hardware, and Ling-2.6-1T improves token efficiency by about 4 times over its predecessor on the Artificial Analysis Intelligence Index. Ring-2.6-1T high scores 87.60 on PinchBench and 63.82 on ClawEval. All three models have been open-sourced.

Key Takeaways

  • Ant Ling released the Ling & Ring 2.
  • 6 technical report,systematically disclosing the architecture,pre-training,post-training and agent…
  • 6-flash,Ling-2.
  • 6-1T and Ring-2.
  • The three models adopt a Hybrid Linear Attention architecture that combines Lightning Attention…

Quick Facts

Published
Updated
Source
公众号:蚂蚁百灵(Ling)
Category
Model Releases
AI score
79
Read the original Read the full article at the original source
Why it matters

蚂蚁百灵2.6技术报告首次公开Hybrid Linear Attention与KPop Agent RL细节,开源模型在OpenClaw登顶,把万亿模型从聊天拉到真实工作流,做Agent应用的值得细读。

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.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: Ant Ling Releases Ling & Ring 2.6 Technical Report. June 16, 2026. https://new.nat6.net/en/news/cmqg1bu4n02o9slsp8813ls12/

View all →