AI Trends

UCSD Hao AI Lab Releases Zero-Shot Robotic Manipulation Framework ORA v0.1

UCSD Hao AI LabUpdated Nov 19 09:08

AI summary

UCSD Hao AI Lab has released ORA v0.1, an open zero-shot robotic manipulation framework that increases the RoboDojo completion rate on DeepSeek V4.1-Flash from 5.2% to 39.9% and the RoboTwin completion rate from 71.5% to 79.6%. ORA enables models to write tools on the fly and uses cuRobo to plan actions based on depth maps; the team also released ORA Traces, comprising 7 models, 5 benchmarks, and 3K+ episodes.

Why read this

AI score 67

The materials also provide cross-benchmark results, code patterns, and changes in collision rates, enabling comparisons of the engineering trade-offs among open robotic agent frameworks.

Topics

Tags

Body

Releasing ORA v0.1, our open harness for zero-shot robot use!

On DeepSeek V4.1-Flash, compared with ORA October’s preview: - RoboDojo: 5.2% → 39.9%, - RoboTwin: 71.5% → 79.6%.

🌐 Code https://t.co/M3ZAFAZOs9 https://t.co/YaDpxEQUV2

Source

UCSD Hao AI Lab