GPT-6 Astra Generally Available

Discussion (6 Comments)Read Original on HackerNews

rc1•about 1 hour ago

Quite hidden in the mobile app at least. I had to click thinking. The slider goes 5.6 Instant, 5.6 Medium, 5.6 High, 5.6 Extra High, and the surprise 6 Pro

minimaxir•about 1 hour ago

I see this as well. Odd there is only 6 Pro.

minimaxir•about 2 hours ago

Also is rolling out to the $100/$200 Codex subscriptions, I got access this morning.

Unsurprisingly, it does indeed consume usage at ~2.5x the rate of Sol.

pllbnk•about 2 hours ago

Shouldn't this new reduction in thinking output make the model cheaper to operate? Could it be that the model is the same old LLM, with a bit newer architecture but still doing the same things, including huge amounts of thinking, just that its thinking is less clear to human readers?

vessenes•33 minutes ago

Those tend to be under so-called ‘max’ modes, or ‘ultra’ - where you scale up inference time compute and then choose amongst your answers. Astra is new weights.

minimaxir•about 1 hour ago

Thinking output is already a relatively small part of the total input compared to raw code/text files.

GPT-6 Astra Generally Available

⚡ Community Insights

Discussion (6 Comments)Read Original on HackerNews