Chenxiao Gao

This account is verified. Learn more

@chenxiaogao

about.gaocx.io | Ph.D student @GeorgiaTech | Prev @Kimi_Moonshot
Foundation models / agentic systems for robotics

Joined February 2024

95 Following 87 Followers 20 Posts

Latest posts 5

Chenxiao Gao @chenxiaogao 21h

Presenting RLE-Bench at #CoRL2026 Agentic Robotics Workshop today:
📍 JW Marriott Austin, Rooms 203–204
🕓 Poster session: 4:20–5:55 PM CT

rlebench Verified account @rlebench 22h

🚀 The RLE-Bench leaderboard finally has a new #1!

Claude Fable 5.2 takes the top spot 🏆, leading 5 out of 9 robot learning engineering tasks! 🤖

Curious how frontier coding agents perform in robotics?

Come chat with us at the #CoRL2026 Agentic Robotics Workshop TODAY!

📍 JW Marriott Austin, Rooms 203–204
🕓 Poster session: 4:20–5:55 PM CT

🔗 rle-bench.github.io

View quoted post on X

Chenxiao Gao @chenxiaogao 5d

Just landed in Austin! I’ll be at CoRL Nov 8–12, presenting RLE-Bench at the CoRL 2026 Workshop on Agentic Robotics. Would love to chat about:
- Benchmark construction
- Agent harness for robotics
- Adapting VLAs/WBCs for agentic systems
- Agent training; and beyond.

DM me if you’d like to chat! Some of our recent work ⬇️
- RLE-Bench: rle-bench.github.io
- Kinex, the harness that connects your coding agent to the robots: kinex-agent.org

Chenxiao Gao @chenxiaogao Nov 5

SuperMimic gets lots of attention for efficient motion tracking RL training.

A complementary question: what should a controller track to make tracking easier while giving high level planners precise and steerable control?

Check out our paper 👇

Feiyang Wu Verified account @fei_yang_wu Nov 5

Humans learn how to solve tasks by composing what they have learned. Can robots do the same? Can robots walk while squatting once they learn individual motions?

Introducing Spectral Skills. Steer, mix, and chain motion skills through composing energy functions.🧵

View quoted post on X

Chenxiao Gao @chenxiaogao Nov 3

K3.1 Pro is officially my go-to model for generating robotics task trajectories

rlebench Verified account @rlebench Nov 3

🚀 Open-weight models are catching up fast in robotics!

Kimi K3.1 Pro just landed at #7 overall and #1 among open-weight models we tested on RLE-Bench!

This is another exciting milestone for open-weight AI. The gap between open and proprietary frontier models is narrowing on complex robotics tasks that require physical reasoning and experimentation. Exciting to see more capable agents becoming accessible to the broader robotics research community.

🔗 Full leaderboard: rle-bench.github.io

#RLEBench #Robotics #OpenSource #Kimi #EmbodiedAI

View quoted post on X

Chenxiao Gao @chenxiaogao Oct 29

🚀 Frontier models are becoming increasingly capable in robotics!

RLE-Bench has also been accepted to the CoRL Workshop on Agentic Robotics! Would be fun to chat if you are interested in benchmarks/harnesses/agent training/new opportunities for foundation models in robotics.

agentic-robotics-workshop.github.io

rlebench Verified account @rlebench Oct 29

🚀 Gemini 4 Argon takes #1 on 4 robotics tasks in RLE-Bench!

The latest Gemini 4 Argon results are in: 72.4 RLE Index, ranking #2 overall, just behind GPT-6 Astra (73.4).

More impressively, Gemini 4 Argon takes the top spot on 4 out of 9 tasks:

* Harness Engineering
* Embodied Reasoning
* Whole-Body Motion Tracking
* NanoVLA Recipe

It’s exciting to see frontier AI labs continuing to invest in robotics capabilities, with each new generation pushing the boundaries of what general-purpose AI agents can accomplish in embodied and physical-world tasks.

The competition is getting closer, and the progress is accelerating.

Looking forward to the Carbon! 👀

Explore the full leaderboard: rle-bench.github.io

#Robotics #EmbodiedAI #RLEBench #Gemini4

View quoted post on X