about.gaocx.io | Ph.D student @ @GeorgiaTech | Prev @Kimi_Moonshot Foundation models / agentic systems for robotics
- Tweets 11
- Following 66
- Followers 39
- Likes 38
Chenxiao Gao retweeted
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.
🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.
Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
1,249
3,231
30,140
4,026,056
Introducing Kinex, the harness that connects coding agents to any physical robot, unlocking scalable learning in the physical world.
Tell Kinex about your robot; it then carries out tasks and learns from every run. We demonstrate Kinex across diverse real-world applications in a zero-shot setting, from bimanual manipulation to quadruped robots.
🧵1/5
3
11
25
71,781
More details and demos on our project website:
stacx-lab.github.io/kinex/
Try it out and see how coding agents can control real robots 🤖
1
144
Kinex decomposes a robot into a collection of devices, and the agent interacts with devices through device drivers, tools, and under the guidance of skills.
🧵2/5
1
1
2
481
Try Kinex in simulation or on your own robot.
Install → run Kinex → connect a model → set up your devices → describe the task.
🧵4/5
1
1
2
200
Kinex is built on the Pi harness.
Plugins also bring Kinex to Pi, Codex, and Claude Code.
Check out the docs, demos, and code here: stacx-lab.github.io/kinex/
🧵5/5
1
3
184
New on the RLE-Bench leaderboard 🚀
Frontier models keep getting better at robotics, and RLE-Bench gives us a measurable way to track it.
Two updates on RLE-Bench! 🤖
📈 New results for GPT-6.1-Sol, Opus 5.5, Fable 5.1, GPT-6-Luna, and Grok 4.6! Opus 5.5 shows substantial gains in interactive control—even outperforming Fable 5.1. Exciting to see frontier models making progress on robotics tasks!
📑 Our technical report is out! We dive into the design principles behind RLE-Bench, break down each task, and share observations of how agents tackle them. Plenty of interesting agent behaviors to explore!
Check out the latest results and read the report 👇
1
2
181
Excited to share RLE-Bench, a benchmark we’ve been building over the past few months. We envision agents that can perceive, control robots, develop policies, and design their own hardware. RLE-Bench measures how close today’s coding agents are to that vision.
GPT-6 Astra is impressive, but agentic robotics ≠ just control.
Physical agents should control robots, learn new skills, and design their own hardware.
So we introduce RLE-Bench: 48 everyday robotics engineering tasks spanning closed-loop control, policy learning, perception, and mechanical design.
2
10
612