GPT-6.2 Astra and Sol: Prices, Plans and What Changes
GPT-6.2 Astra and Sol launched 17 November 2026 in ChatGPT Work, Codex and the API. Prices, plans, usage limits and who should switch, from OpenAI's own docs.
PromptWises is reader-supported. Some links may earn us a commission at no extra cost to you. This never affects our ratings or recommendations.
OpenAI released GPT-6.2 Astra and GPT-6.2 Sol on 17 November 2026 in ChatGPT Work, Codex and the API. Astra is the model for the hardest projects; Sol is the everyday model for coding, writing and multi-step work. If you pay for ChatGPT Plus or above, Sol is the new default to try in Work and Codex. If you build on the API, Astra is now cheaper per token than GPT-6 Astra, and Sol costs the same as GPT-6.1 Sol. OpenAI’s announcement page was not readable when we checked, so every fact here comes from the ChatGPT launch note, the API changelog, the GPT-6.2 Astra and GPT-6.2 Sol model pages and the ChatGPT pricing documentation.
What changes in ChatGPT and Codex
Both models are available in ChatGPT Work on the web, desktop and mobile, and in the Codex app and CLI, on Plus, Pro, Business, Enterprise and Edu. They are not on Free or Go, and they are not in Chat: OpenAI says to open Work to use them. Enterprise and Edu admins must switch the models on in workspace settings first.
New Work chats on consumer plans default to Sol at Light reasoning effort; Enterprise workspaces default to Sol at Medium. Work and Codex share one usage allowance. OpenAI’s estimates for Plus and Standard Business, in local messages per five hours:
| Model | Plus / Standard Business (per 5 hours) |
|---|---|
| GPT-6.2 Astra | 6-50 |
| GPT-6 Astra | 5-45 |
| GPT-6.2 Sol | 17-180 |
| GPT-6.1 Sol | 15-160 |
So the new models stretch a Plus allowance slightly further than the ones they replace. OpenAI’s launch note says Sol “lets you complete more tasks on the same paid plan” than GPT-6.1 Sol. Pro plans still have no five-hour limit. Ultrafast works with Sol on Pro $200, and with both models on Pro $500 and eligible Enterprise and Edu workspaces once an owner enables it. For plan prices, see our ChatGPT review.
API pricing and limits
| Per 1M tokens (Standard) | GPT-6.2 Astra | GPT-6 Astra | GPT-6.2 Sol | GPT-6.1 Sol |
|---|---|---|---|---|
| Input | $8 | $10 | $2 | $2 |
| Cached input | $0.40 | $1 | $0.10 | $0.10 |
| Output | $40 | $50 | $10 | $10 |
Prices are at the time of writing. Both models have a 1,050,000-token context window (922,000 maximum input) and 128,000 maximum output tokens. They take text and images and return text. Knowledge cutoffs are 18 May 2026 for Astra and 1 June 2026 for Sol. The fine print matters for big jobs:
- Requests over 272K input tokens pay double for input and cache, and 1.5x for output, on the whole request.
- Batch is 50% of Standard; Fast mode is 2x.
- Ultrafast is 3x Standard until 28 February 2027, then 6x from 1 March 2027.
- Tool calling needs the Responses API, and the API rejects reasoning effort
none. - US and EU data residency are supported; Ultrafast is US-only.
The same day OpenAI opened a background completion window beta in the Responses API: a two-hour window is 30% cheaper than Standard and an eight-hour window 40% cheaper, for GPT-6.2, GPT-6.1 Terra and GPT-6 Luna. For overnight reports and bulk processing, that is the cheapest way to run the new models short of Batch.
GitHub Copilot gets it the same day
GitHub made both models generally available in Copilot on Pro+, Max, Business and Enterprise, across VS Code, Visual Studio, JetBrains, Xcode, Eclipse, the Copilot CLI, the coding agent and github.com, with a gradual rollout. They are billed at OpenAI’s list price under usage-based billing. GitHub says that in its own early testing GPT-6.2 Sol caught nearly twice as many real issues in code review as GPT-6.1 Sol. That is GitHub’s claim, not an independent benchmark. If you are weighing editors, our Cursor review covers how Cursor prices access to third-party models.
What to do now
- ChatGPT Plus users: nothing to change. Work already defaults to Sol. Pick Astra only for long, multi-source or multi-service jobs, because it uses your allowance about three times faster.
- API teams on GPT-6 Astra: test GPT-6.2 Astra. It is 20% cheaper per token. Check the GPT-6.2 migration guide first, because
nonereasoning effort is rejected. - API teams on GPT-6.1 Sol: same price, so the switch comes down to quality on your own evaluations.
- Enterprise admins: the models stay off until you enable them.
OpenAI has not announced a retirement date for GPT-6 Astra or GPT-6.1 Sol. We track shutdowns in our Q4 2026 model retirements tracker, and our guide on how to read AI model announcements explains why vendor benchmark claims deserve a second look. More launches are on the news page.
Frequently asked questions
How much does GPT-6.2 cost in the API?
At the time of writing, GPT-6.2 Astra costs $8 per million input tokens, $0.40 cached input and $40 output. GPT-6.2 Sol costs $2, $0.10 and $10. Requests with more than 272K input tokens pay double for input and 1.5x for output, and Batch costs half the Standard rate.
Can I use GPT-6.2 in ChatGPT's normal chat?
No. OpenAI's launch note says that to use GPT-6.2 Astra or Sol in ChatGPT you open Work. The models you can pick in Chat did not change. Both models are also in the Codex app and CLI.
Is GPT-6.2 available on the free ChatGPT plan?
No. GPT-6.2 Astra and Sol are available on Plus, Pro, Business, Enterprise and Edu. In Enterprise and Edu workspaces they are off by default until an administrator turns them on.
Should I use GPT-6.2 Astra or GPT-6.2 Sol?
OpenAI positions Astra for the hardest work, such as tracing a bug across several services or comparing many research sources, and Sol for everyday coding, writing and multi-step tasks. Sol costs a quarter of Astra per token and uses much less of a Plus plan's Work and Codex allowance, so start with Sol and move up only when it falls short.
Is GPT-6.2 in GitHub Copilot?
Yes. GitHub made both models generally available on 17 November 2026 for Copilot Pro+, Max, Business and Enterprise, billed at OpenAI's list price under usage-based billing. Copilot Pro is not on GitHub's list.