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.

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:

Oracle killed it w/ a big beat. Coming off the back of GPT6 Astra too I expect them to drift higher. Oracle might be a viable neocloud, all the channel checks were terrible but I just think it’s cause Larry Ellison is a psychopath who extends his credit too far with suppliers.

Blew through both my Claude Max 20x and ChatGPT Pro 20x weekly limits in ~2 days using Fable 5.1 + GPT6. Some observations: 1. Claude Max 20x ran out noticably faster for me than ChatGPT Pro 20x using Fable 5.1 and GPT6 respectively 2. Fable 5.1 writes much better plans for humans to read (qualitative). Both one shot my work in ~90% cases though 3. Claude Code TUI >> Codex CLI in terms of product craft & user experience 4. Codex Mac app computer use is excellent. It does a ton of rote tasks for me that I didn't trust models to handle before Huge jump in capabilities, curious to see how (4) impacts jobs that do a lot of repetitive tasks on the computer

ICLR 2027 submission: we benchmarked all frontier models and our model performs... Anthropic & OpenAI today: no you don't. ICLR deadline is 9/25, so i guess no excuse not including opus5.5 and gpt6-luna & sol in the paper.

Who's better at making movies? GPT 6 Astra VS Deepseek 4.1 Flash making AI videos (Documentary / Wizard Movie / Fight Scene / Travel Vlog) Create the same videos with templates & prompts ready at @FramiaPro gpt6astra is available at Framia Agent!

See a viral ad you love? Recreate it in one click with Topview AI Marketer! Keep the viral formula while swapping in your own products seamlessly #TopviewAI# #AIMarketer# #AIVideo# #EcommerceMarketing# #GPT6#