The development
What happened
OpenAI released two GPT-6.2 reasoning models on November 17. It designed Astra for complex reasoning and professional projects, and Sol for everyday coding, writing and tasks that take several steps. Both are available in ChatGPT Work, Codex and the OpenAI API, with access depending on the plan and workspace settings.
Key facts
OpenAI lists ChatGPT Work access on Plus, Pro, Business, Enterprise and Edu. Enterprise and Edu administrators must first enable the models. In Codex, users can choose a model in the app or CLI.
Developers can select gpt-6.2-astra or gpt-6.2-sol in the Responses API. OpenAI lists U.S. and EU data residency for Standard, Fast and Batch processing. Ultrafast is limited to U.S. data residency.
In Work and Codex, OpenAI’s guide lists Ultrafast for Sol on the $200 Pro plan and for both models on the $500 Pro plan. Eligible Enterprise and Edu workspaces can use it after an owner enables it.
At Standard API rates for requests with up to 272,000 input tokens, Astra costs $8 per million uncached input tokens and $40 per million output tokens. Sol costs $2 and $10, respectively.
The context
Why it matters
TideReach assessment. The two models give teams a higher-priced option for complex work and a lower-priced option for routine tasks. Astra’s published Standard rates are four times Sol’s for the same number of uncached input or output tokens. The cost of a completed task also depends on how many tokens each model uses.