Tian Tang @ ICML 🇰🇷

MTS @NexEcosystem
B.E. YaoClass @Tsinghua_IIIS @Tsinghua_Uni
Ex @uwcse | All opinions are my own

[Tian Tang @ ICML 🇰🇷 @Tangt03]

I’m at #ICML2026 this week! Let’s connect, grab a coffee, or talk shop. ☕️
Check out our poster session: 🗓️ July 7 @ 10:30 AM | Hall A #3913
🚀 Hiring Alert: Looking for talented folks interested in LLM & Agent jobs/internships. Drop me a DM or find me at the poster!

[Nex @NexEcosystem]

📢 Update: Nex-N2-Pro is now transitioning from free to paid on OpenRouter.
🙇 Thanks to everyone who tried it during the free period! The model ID has been updated:
👉 nex-agi/nex-n2-pro:free → nex-agi/nex-n2-pro

💰 Pricing:
· prompt tokens: $0.25 /1M
· cached prompt tokens: $0.025 / 1M
· generation tokens: $1.0 / 1M

We’d love for you to keep using it and share your feedback — it helps us improve.
https://t.co/QOrT9jJ4CL

[Tian Tang @ ICML 🇰🇷 @Tangt03]

@SakshiSugandhi gpt 5.5

[Tiezhen WANG @Xianbao_QIAN]

wait... what???

https://t.co/NxKCkU2Qnr

[Max For AI @MaxForAI]

卧槽？？

昨天爆火的巴西Rio开源模型，被抓到原来是套壳Nex-N2了🤣

昨天爆火的由巴西里约热内卢市政府旗下 IT 公司开源的模型 Rio 3.5 397B。

被GitHub的网友发现去掉 Rio 硬编码的 “You are Rio” 系统提示词之后，它自己部署的模型有 79% 的时候会把自己识别为：
“Nex，来自 Nex-AGI”

而识别为 “Rio” 的比例是 0%。

它���至会逐字复述Nex-N2专门写的背景故事。

Rio 里的每一个权重张量，在统计上都表现为同一个 0.6/0.4 的 Nex 和 Qwen 混合比例。这个现象覆盖全部 60 层，以及网络中的每一个组件。

正常的其他微调模型，无法被解释成这种线性插值结果。

微调一般是拿 Qwen 当底座，用数据继续训练，通过梯度下降更新权重，得到一个新模型

这只能是用的现成的权重了😂

[Nex @NexEcosystem]

The Rio 3.5 model broke the internet this week. The plot twist? It’s essentially our open-source model, Nex N2 Pro, wearing a different hat.

🤯 We analyzed the weights, and the recipe is exact: Rio 3.5 ≈ 0.6 * Nex N2 Pro + 0.4 * Qwen 3.5

It even literally introduces itself as "Nex N2 Pro" if you ask it without initial system prompt!

😂 We are flattered that the City of Rio used our work to achieve SOTA performance. Thanks for the ultimate benchmark validation.

🤝 But in the open-source world, attribution matters.
👇 Full mathematical proof & verify script in the first reply!

[Tian Tang @ ICML 🇰🇷 @Tangt03]

A hexagon brick game built with Nex-N2-Pro — smooth, polished, and seriously impressive to look at! Nex-N2-Pro is now free to use on OpenRouter, try it out! https://t.co/C8GDkLjtlM

[0xSero @0xSero]

Really good model, enjoyed using it. Feels like a complete package.

[Tian Tang @ ICML 🇰🇷 @Tangt03]

@uzairansar Maybe you should check out our new model: https://t.co/xNX7yGV2Du

[Nex @NexEcosystem]

🔥 Nex-N2-Pro is now live on OpenRouter — completely FREE to try!
Give it a spin: https://t.co/nCA5D1WOYx
Drop your results below ⬇️

[David Hendrickson @TeksEdge]

🔥Holy 💩! New King of the Open Source LLMs! Beats Leading OS Models. Time to start testing to verify benchmarks.

🏆 Nex-N2-Pro beats
✅ Kimi-K2.6: 86% (13/14)
✅ MiniMax: 100% (8/8)
✅ GLM-5.1: 100% (13/13)
✅ DeepSeek-V4-Pro: 92% (11/12)
» Opus 4.7: 45% (5/11)
» GPT-5.5: 30% (3/10)

[ModelScope @ModelScope2022]

Nex-N2 is now open source！An agentic model series from Nex AGI built for coding, tool use, deep research, and long-horizon workflows. 🧠🔎

🛠️ https://t.co/p2n2cMdBlR
⚙️ https://t.co/k67uqVgDDP

● Models: Nex-N2-Pro 397B total, 17B active; Nex-N2-mini 35B total, 3B active

● Agentic Thinking: adaptive reasoning depth + coherent reasoning across coding, search, tool calling, and execution

● Efficiency: Nex-N2-mini saves roughly 20% overall token cost vs forced thinking while matching or slightly exceeding task performance

● Open-model lead: 75.3 on Terminal-Bench 2.1, 80.8 on SWE-Bench Verified, 83.7 on BrowseComp, and 1585 on GDPval among listed open baselines

● Deployment: customized SGLang fork, reasoning parser, tool-call parser, Docker image

● License: Apache 2.0

[Hugging Models @HuggingModels]

Meet Nex-N2 Pro, a cutting edge MoE model that handles both images and text. It's designed for complex conversations and generation tasks, making it a versatile tool for developers and creators. This is the future of multimodal AI.

[Nex @NexEcosystem]

Nex-N2-Pro now available on @novita_labs! Try it out if you're building coding agents or research workflows 🚀

[Tian Tang @ ICML 🇰🇷 @Tangt03]

Try out our new opensource model for free! Performance is close to Opus 4.7!

[Dawning Road @TheDawningRoad]

Introducing Nex-N2 — true Agentic Thinking, built with @NexEcosystem 🚀

Thinking is now standard in foundation models, but it sits in an awkward position in Agent tasks: either the performance gains aren't significant, or it's verbose, and switching scenarios means readapting all over again. The root cause is that thinking from the o1/R1 era was built around RLVR for math and code tasks, not for long-horizon Agent tasks—there's a layer of separation between thinking and action.

Nex-N2 introduces a complete Agentic Thinking framework, split into two parts: Adaptive Thinking and Coherent Thinking. The former achieves adaptive reasoning intensity, improving speed (which really matters in long-horizon tasks spanning hundreds of steps) and saving unnecessary token expenditure. The latter unifies thinking patterns across different tasks, making actions more stable, consistent, and robust.

- Adaptive Thinking, auto-scales reasoning depth per step. Saves ~20% tokens, zero performance loss.
- Coherent Thinking, one thinking paradigm across search, coding, and tool use. No more fragile mode-switching.

On coding and Agent tasks, Nex-N2 ranks in the top tier of open-source models. The model is fully open-sourced and available simultaneously on Hugging Face, ModelScope, and SiliconFlow. We welcome everyone to try it out.

Official website: https://t.co/8j4ZvLAQZt
Huggingface: https://t.co/UEx20Th3wQ

[Tian Tang @ ICML 🇰🇷 @Tangt03]

Best open source coding model! Try it for free!

[Nex @NexEcosystem]

📢 Nex-N2 is here!
A family of agentic models that doesn't just think, it acts!
Coding, search, tool use. All fused into a single agentic reasoning loop.

- Adaptive Thinking, auto-scales reasoning depth per step. Saves ~20% tokens, zero performance loss.
- Coherent Thinking, one thinking paradigm across search, coding, and tool use. No more fragile mode-switching.

🏆 Result: Tier-1 open-source performance on SWE-bench, Terminal-Bench, GDPval, and more, tracking GPT-5.5 and Opus 4.7.

🎉 Open-weight. Try it now.
🔗 https://t.co/7oLSfyOCxB
📦 https://t.co/c2CGhXWaz6
https://t.co/KJYXZIpk8M
https://t.co/vcjdZ9cuB6

[Tian Tang @ ICML 🇰🇷 @Tangt03]

@QuanquanGu All the best!

[Tian Tang @ ICML 🇰🇷 @Tangt03]

@karthikponna19 Rust, Go

[Tian Tang @ ICML 🇰🇷 @Tangt03]

Codex is dumb today, not even remembering why one file was changed, what happened?