Give two people the same problem.
Give them the same AI tools, the same models and unlimited tokens.
One builds something functional. The other builds something extraordinary.
What separates them?
An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.
We believe it will be a major step for scientific reasoning. openai.com/index/ten-adva…
The same agent job — 10M tokens in, 1M out — costs $1.68 or $150 depending on who you buy it from.
An 89× spread across models that score within 24 points of each other.
And the second-best intelligence-per-dollar in that whole spread isn’t an open-weight Chinese model. It’s
OpenAI cut GPT-5.6 Luna 80% yesterday.
A 10M-token agent run: $3.20 on Luna, $150 on Claude Fable 5.
Luna scores 51 on Artificial Analysis’ Intelligence Index — 10 points off the frontier — at roughly 1/25th the blended cost of GPT-5.6 Sol.
Cost has stopped being the reason to
we've cut prices on luna by 80%, making it by far the most price-efficient model in its class.
a lot of our research is about how to create incredibly efficient models for any given level of intelligence.
excited to see what you all do with intelligence too cheap to meter!
Everyone’s talking about AI-native interfaces.
We’ve actually built one.
The best way to understand it isn’t by watching the video—it’s by trying it.
Download ixigo NEXT, talk to TARA, and let me know what surprises you.
app.adjust.com/18bmuyau