Zhiheng Xi

PhD Candidate at Fudan NLP Group of @FudanUniversity. Prev. BS at @NJU1902. @NexEcosystem @TencentHunyuan @ByteDanceSeed @Meituan_LongCat

[Zhiheng Xi @Be1ong1]

🚀 Excited to introduce AI for Productivity in the Age of Agentic AI — a 68-page report I co-led with the AI4X Project Team, bringing together researchers from 12 academic institutions worldwide, including @FudanUniversity, @ZJU_China , @Stanford, @UCBerkeley, @UniofOxford , and @Princeton.

📄Paper: https://t.co/qDgIFvtQ5N
💻Repo: https://t.co/WIZwBDwDhG

🕶️In the age of agents, AI is becoming a new productive instrument. Its significance goes beyond making individual tasks faster. Agentic AI is beginning to reorganize how work is executed, coordinated, and delegated — and, in turn, reshape the relationship between humans, organizations, and intelligent systems.

🧐This leads to the central question of our report: How does AI capability actually translate into sustained productivity growth?

#AgenticAI #AI #Productivity #AI4Productivity

[Zhiheng Xi @Be1ong1]

🧐Our key argument is that the next wave of AI-driven productivity will not come only from doing the same work faster. It will also come from changing what work can be done, how it is organized, and how responsibility is distributed between humans and AI. This report is our attempt to systematically map that transition — and to understand what the next paradigm of production may look like.

[Zhiheng Xi @Be1ong1]

🚀 Excited to introduce AI for Productivity in the Age of Agentic AI — a 68-page report I co-led with the AI4X Project Team, bringing together researchers from 12 academic institutions worldwide, including @FudanUniversity, @ZJU_China , @Stanford, @UCBerkeley, @UniofOxford , and @Princeton.

📄Paper: https://t.co/qDgIFvtQ5N
💻Repo: https://t.co/WIZwBDwDhG

🕶️In the age of agents, AI is becoming a new productive instrument. Its significance goes beyond making individual tasks faster. Agentic AI is beginning to reorganize how work is executed, coordinated, and delegated — and, in turn, reshape the relationship between humans, organizations, and intelligent systems.

🧐This leads to the central question of our report: How does AI capability actually translate into sustained productivity growth?

#AgenticAI #AI #Productivity #AI4Productivity

[Zhiheng Xi @Be1ong1]

We take a cross-industry view of how Agentic AI is diffusing into real production workflows.

[Zhiheng Xi @Be1ong1]

🧐Our key argument is that the next wave of AI-driven productivity will not come only from doing the same work faster. It will also come from changing what work can be done, how it is organized, and how responsibility is distributed between humans and AI.

This report is our attempt to systematically map that transition — and to understand what the next paradigm of production may look like.

[Zhiheng Xi @Be1ong1]

We take a cross-industry view of how Agentic AI is diffusing into real production workflows.

[Tencent Hy @TencentHunyuan]

🚀Hy3 is here.

295B MoE. Best in its size class. Rivals trillion-scale flagships.
Reliable and affordable for most agentic usecases.
Apache 2.0. Friendly for commercial use.
FREE API for 2 weeks → https://t.co/EyURKwTdgi

🤗 https://t.co/twqJpqb2SL
📖 https://t.co/4uEkIU1cW4

[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

[Zhiheng Xi @Be1ong1]

I’m here at #ICML2026 🇰🇷with my labmates!

Feel free to come say hi and hang out!
(Photo 2 shows the wonderfully absurd room that the platform accidentally booked for me.)

My presentations are as follows:
1. RL generalization research for LLM agents
2. SciAgentGym: research on scientific agents
3. Scalable Interactive Oversight: scalable interactive supervision for agents
4. ChartE3: a new benchmark for text-image editing

[Zhiheng Xi @Be1ong1]

Great Memories in Rio！

[Zhiheng Xi @Be1ong1]

🤩 My most extroverted moment at Oral presentation at ICLR 2026

[Zhiheng Xi @Be1ong1]

I’ve landed in Rio! My upcoming presentation schedule is below. Feel free to come find me to hang out, chat, or grab a meal!

[Zhiheng Xi @Be1ong1]

@ScalerSohom 🥰

[Zhiheng Xi @Be1ong1]

Welcome, to our new work, BAPO, which tackles the challenge of Off-Policy Reinforcement Learning for Large Language Models across settings like Partial Rollout and Experience Reuse.

Code: https://t.co/ULtuq04vpE
Paper Link: https://t.co/oS5qxAhaRq