Oct 8, 2026 · 8 min · 6 segments
Thursday morning AI and cyber briefing. OpenAI moves everyday ChatGPT onto GPT-6, with Sol for paying users and Luna for free users, plus Intelligent UI answers built from charts, buttons and small…
Yesterday, it began putting them into the everyday ChatGPT chat window, which more than a billion people use every week.
Paying users get GPT-6 Sol, and from today, free users get Luna, a smaller and cheaper version.
The headline feature is called Intelligent UI, where UI stands for user interface, the part of the screen you actually use.
Instead of only writing text, the model can build an answer out of charts, diagrams, buttons, and forms, and even make a small working tool on the spot, like a savings calculator.
OpenAI's safety report, which it calls a system cart, says the new models resist tricks to get round their rules better than before, but they're also more willing to engage with borderline requests, and they slipped on several tests for keeping teenagers safe.
For business accounts, it only switches on when the workplace administrator allows it.
This one's for Tim because staff will notice the change before IT does.
ChatGPT is changing under people's feet with a new model and answers that look more like little apps than paragraphs.
Most of that is good for users, but when an answer looks polished, people tend to trust it more and paste it straight into their work.
If you run ChatGPT for a business, check that administrator setting today and choose when it goes live rather than letting it arrive by surprise.
And if you look after schools or younger users, read the teen safety section of that report before you roll it out.
A new look, a new model, and one setting to check first.
Yesterday, it began putting them into the everyday ChatGPT chat window, which more than a billion people use every week.
Paying users get GPT-6 Sol, and from today, free users get Luna, a smaller and cheaper version.
The headline feature is called Intelligent UI, where UI stands for user interface, the part of the screen you actually use.
Instead of only writing text, the model can build an answer out of charts, diagrams, buttons, and forms, and even make a small working tool on the spot, like a savings calculator.
OpenAI's safety report, which it calls a system cart, says the new models resist tricks to get round their rules better than before, but they're also more willing to engage with borderline requests, and they slipped on several tests for keeping teenagers safe.
For business accounts, it only switches on when the workplace administrator allows it.
This one's for Tim because staff will notice the change before IT does.
ChatGPT is changing under people's feet with a new model and answers that look more like little apps than paragraphs.
Most of that is good for users, but when an answer looks polished, people tend to trust it more and paste it straight into their work.
If you run ChatGPT for a business, check that administrator setting today and choose when it goes live rather than letting it arrive by surprise.
And if you look after schools or younger users, read the teen safety section of that report before you roll it out.
A new look, a new model, and one setting to check first.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
