Background
The GPT-6.2 announcement was published on OpenAI's site and was paired with a statement from Microsoft about the same model.
What happened
On 17 November 2026, OpenAI announced GPT-6.2 Astra and Sol, with the announcement timestamped 17:45 UTC.[1] The announcement covers ChatGPT Work, Codex and the API, the three places where OpenAI says the models are available.[1]
Key details
OpenAI's stated focus is professional outputs, coding, computer use and lower-cost efficiency.[1] Computer use means the model operates software the way a person would. Efficiency at lower cost means the same kind of work can be done for less.
Microsoft said at its Ignite 2026 event that GPT 6.2 would come to Foundry, and described it as 'a great model that sets a new bar for AI in professional work.'[2]
Impact
The release gives people who use ChatGPT, Codex or the API one more model to choose from, with a stated emphasis on work tasks. Microsoft's plan to offer the model through Foundry means the same model will reach customers who build on Microsoft's platform as well. The practical result for most readers will depend on how the models perform on their own tasks, which is something the announcement itself does not measure.
Sources
- 1
- 2
Cite this page: OpenAI releases GPT-6.2 Astra and Sol. TellWorld, https://www.tellworld.net/news/gpt-6-2-2026. Updated 2026-11-18