GPT6 Astra is pretty good — After a morning of intensive coding. Clear improvements over 5.6 Sol, much better frontend taste, much better engineering rigorous and much much more efficient thinking and answering.

ALERT 🚨🚨OpenAI's latest GPT6.1 Sol Ultrafast is NOT running on Cerebras but is instead running at a low batch size on NVIDIA GPUs. What does this say about Cerebras? Will Cerebras be serving GPT6.1 Sol Ultrafast in the future?

Great work that measures GPT6 Astra as an Embodied Policy by Yu-Mool Shu and Lipxin Zheng. It further proves with numbers that VLA handles dexterous movement, while the General Model oversees supervision, error correction, and replanning. Interesting numbers: For the average performance across 10 tasks in RoboDojo, PI0.5 gets 24%, Astra gets 37% but together they achieves 62%. But for RoboLab, Astra itself gets 98%. As for the contribution, GPT 6 Astra is correcting only 14.4% of the executed control steps and the remaining 85.6% following the actions by PI0.5​. Link:

my takeaway from 3 days of GPT6/Astra research: it’s simultaneously far more capable and efficient than expected, yet still far from good enough. that combination should massively accelerate AI adoption, drive Jevons paradox, and expose how much work remains before these models are dependable across the full scope of real world use cases. hard to imagine a more bullish outcome for AI infrastructure.

🥒 Teaching a #reBot# Arm to slice cucumbers with #GPT6Astra—in# the real world! 🤖 Community developer @box2ai brings #LLM# agents to physical #robot# through #ROS#, In-Context Learning, and self-evolving loops—no hand-crafted control pipeline needed. 🦾 Open-source reBot Arm: