Richard Bian

Model as Product @AntLingAGI. Founder of AntOSS. Seasoned Engineer & Strategist. 10+ yrs of eng exp + MBA. Global citizen lived in China, Singapore, Canada, US.

[Richard Bian @RichSFO]

@whatthelukh @ZhipuAI @deepseek_ai @minimaxAI @PrimeIntellect @nvidia @poolside_ai @AIatMeta Thanks Luke for featuring the Ring-1T model! Our official account is @AntLingAGI for reference purpose :) Stay tuned for our technical report!

[Ant Ling @AntLingAGI]

From IcePop to KPop — our team keeps pushing on RL training stability for large MoE models. 👇

KPop replaces the fixed-ratio mask with an adaptive binary-KL region that matches each token's inherent noise. More robust updates, stable long-horizon agentic RL.

Ring-2.6-1T → 76+ on SWE-bench Verified, pure RL.

Congrats to @Jia__Guo & team!

Blog: https://t.co/PvYHN5ywwz

[Richard Bian @RichSFO]

@lifesinger 你说的这两个人是不是一个人 😛😛

[Richard Bian @RichSFO]

@natolambert Can't agree more with your second half statement.

[Richard Bian @RichSFO]

@lifesinger 我很喜欢的一位演讲家 Simon Sinek 有个经典的 leadership 演讲是 Starts with why, people don't buy what you do, they buy why you do it. 私以为然。
做事先做人

[Richard Bian @RichSFO]

IMH the model shouldn't be invisible, it is a kernel or core component or a brand, like "Intel Inside".
If that's a product depends on your perception, or position. To application and harness builder, the model is indeed a product, as it has different behavior and value proposition.

[Richard Bian @RichSFO]

@justone_he 大模型替我读系列，比仓鼠屯书略好一点

[Richard Bian @RichSFO]

@matthew_d_white Ah bummer, we should take a selfie at dinner~

[ModelScope @ModelScope2022]

Ring-2.6-1T is now open source! 1T-parameter thinking model with two reasoning gears and SOTA Agent execution. 🚀
🔗 https://t.co/OI8hBuSJEn

🤖 Stronger Agent execution: SOTA on PinchBench (87.60) and ClawEval (63.82), top-tier on TAU2-Bench and GAIA2-search. SWE-Bench Verified 74.00.
🎯 high / xhigh Reasoning Effort: high for production Agent workflows, xhigh for hard reasoning (AIME 26: 95.83, GPQA Diamond: 88.27)
⚡ Async RL + Popsicle algorithm: stable long-cycle RL training at trillion-scale

Compatible with OpenClaw, Claude Code, OpenCode, Kilo Code, and more.

[Ant Ling @AntLingAGI]

🚀 Ring-2.6-1T is now open source.

A trillion-scale flagship thinking model built for real-world complex tasks: Agent workflows, coding & engineering, long-horizon tasks, complex reasoning, research, and enterprise automation.

It is designed to move beyond “answering” toward execution: understanding context, planning steps, calling tools, and staying stable across long task chains.

Highlights：
- Advanced agentic workflow support.
- Reasoning effort levels: high for agentic tasks, xhigh for complex reasoning.
- Scalable asynchronous RL via the IcePop algorithm, enabling stable, trillion-scale training for long-horizon agentic RL.

[Richard Bian @RichSFO]

It was quite a delight to attend the LLM day at @themu_xyz , a very passionate and energetic event. In my talk, I shared LLM is not a wishing well, and how @AntLingAGI is positioning ourselves in this game. We also had a very informative panel discussion thanks to Xiongyi, Vincent and Jiajun~ A lovely day it is.

[the-mu ✈️ 🇨🇳 @themu_xyz]

We started the Great LLM Meetup yestersay with a stacked docket of talks, panels and fiery discussions 🔥
With guests from Z-ai, Openclaw, Qwen, PixVerse, Kimi, Ant ling Model & StepFun gracing us, we were bound to leave the session feeling like LLM maxis 😎

[Richard Bian @RichSFO]

@ClementDelangue We are open sourcing another model tomorrow! Will tweet against you clem 😃

[Richard Bian @RichSFO]

@themu_xyz Thanks for having me!

[the-mu ✈️ 🇨🇳 @themu_xyz]

We started the Great LLM Meetup yestersay with a stacked docket of talks, panels and fiery discussions 🔥
With guests from Z-ai, Openclaw, Qwen, PixVerse, Kimi, Ant ling Model & StepFun gracing us, we were bound to leave the session feeling like LLM maxis 😎

[OpenRouter @OpenRouter]

In case you missed it: Ring-2.6-1T from @AntLingAGI is live and free on OpenRouter through May 15th!

A 1T-param thinking model with adjustable thinking effort, capable of tackling your production agent needs.

[InfronAI @InfronAI]

Ling-2.6-1T is now available on Infron AI. The brain behind the Flash. Ant Group's trillion-parameter flagship nears GPT-5.4 intelligence at 1/4 the token cost. Try it now 👇
🔗 https://t.co/89DAhEMXfK

#InclusionAI #AntGroup #Ling26 #OpenSource

[Ant Ling @AntLingAGI]

It was quite a delight to get Ant Ling 2.6 1T's Architecture featured in @rasbt 's latest update🤝. We welcome feedback and discussions on our HF page https://t.co/bZwQRsjljU as well as our Discord channel
https://t.co/TwbjGjpDza
Stay tuned for more updates! 🤠

[Richard Bian @RichSFO]

@yihong0618 哈哈哈哈哈哈哈哈会说就多说点

[Richard Bian @RichSFO]

@rasbt Hope you had a nice break Sebastian! Family is very important. 👏

[Richard Bian @RichSFO]

Will be speaking on Tuesday LLM gathering. See you at Monday/Tuesday next week in Shanghai 🥳

[the-mu ✈️ 🇨🇳 @themu_xyz]

https://t.co/dvQPI9LIU3