Vibeleaderboard
← All Intel
Intel / post

OpenAI releases GPT-6.2 Astra and Sol

Source
x.com
Date
OpenAI@OpenAI

Some of you have been asking very politely. Some of you have just been asking. GPT-6.2 Astra and Sol are now available. Super efficient and built to perform where it counts—the real world. https://t.co/qr2iP0GpfX

Context

OpenAI announced that GPT-6.2 Astra and GPT-6.2 Sol are now available, calling them super efficient and built to perform in the real world. The two names mark two roles. In the description of its launch video, OpenAI calls Astra its new frontier model for the hardest projects and Sol its efficient, versatile model for everyday work, names OpenAI also gave to earlier GPT-6 models.

The efficiency claim has a specific meaning in OpenAI's telling. The same description says both models deliver better results on demanding tasks while generating fewer output , the units of text a model writes and is billed for. If that holds, a ChatGPT subscriber gets more done before reaching a usage limit, and a developer paying per token spends less per finished task. Neither the post nor the description gives a figure, and OpenAI's announcement article could not be retrieved (it returned an access error), so any measurements it contains are not reflected here.

The post's video shows what OpenAI means by the real world: office work. In an automatic transcript of its audio, with the visuals not analyzed, employees scramble to update payables from 20 invoices that have just arrived by email, finish a missing presentation slide, and build a workbook modeling a one-year delay in savings on an integration, all while their boss rides up a long escalator toward them. It matches OpenAI's claim that Astra produces slides, spreadsheets and documents that follow a field's conventions more closely, but it is an advertisement rather than a test. What the post establishes is availability. The quality and efficiency claims still need independent measurement, and developer access and pricing are covered in OpenAI's developer announcement.

Terms in this piece · Glossary
  • token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
More from OpenAI
Recommended reads
Comments

Checking sign-in…

Loading comments…