What you will have at the end
By the end of this guide you will know where GPT-6 Astra is available on your plan, you will have selected it for one real task, and you will know how to read its usage before a large job. OpenAI describes Astra in its help center as its most capable model for coding, research, analysis and complex problem-solving.
Before you start
Astra in ChatGPT has two entry points with separate limits, so first check which one your plan includes.
- Chat: GPT-6 Pro, powered by GPT-6 Astra, is available on Pro $100, Pro $200, Business and Enterprise plans.
- Work and Codex: Plus includes Astra here, but not GPT-6 Pro in Chat.
- Enterprise: access also depends on workspace model-access permissions.
- The latest ChatGPT desktop app (Menu > Check for Updates). If you use the Codex CLI, version 0.153.0 or newer.
1. Update the desktop app
Open the ChatGPT desktop app and choose Menu, then Check for Updates; on a Mac this is in the app menu in the top menu bar. Install any update, fully quit and reopen the app, then check for updates once more, because OpenAI notes a second update may be needed even after a recent one.
2. Select Astra in Chat
In the model menu choose GPT-6, then move the Thinking slider to Pro. According to OpenAI, Instant, Medium, High and Extra High run on GPT-6 Sol, while Pro uses GPT-6 Pro, powered by Astra. Intelligent UI, which mixes text with visuals, is not available at Pro effort, so expect plain answers there.
3. Select Astra in Work or Codex
In Work or Codex, open the model picker and choose GPT-6 Astra, then set the reasoning level separately. OpenAI's usage guide says lower effort does not mean lower capability across models and that Astra at Low can outperform Sol at High, so it suggests starting Astra at Low or Medium if Sol at High has worked for you. That is OpenAI's claim, not an AppFor.ai test.
4. Run a first task
Suggested example, not a recorded AppFor.ai run: give Astra one bounded problem you already understand, such as a bug with a known cause, at Medium effort. Ask it to explain the cause before proposing a change.
5. Check the result and your usage
Keep the change only if all of these hold; otherwise give Astra the specific gap rather than raising effort first. OpenAI notes that higher effort cannot supply missing files, access or instructions.
- The explanation matches the cause you already know
- The diff touches only the named file and its test
- The new test fails before the change and passes after it
- Settings, Usage shows how much of your five-hour and weekly allowance the task used
Limits and common problems
Work and Codex share one allowance. For Plus and Standard Business, OpenAI estimates 5 to 45 local Astra messages per five-hour window, against 15 to 160 for GPT-6.1 Sol; these are estimates, and cloud tasks may use more. Pro $100, $200 and $500 currently have no five-hour limit in Work and Codex but still have an included allowance. Chat limits for GPT-6 Pro are separate.
If Astra is missing, confirm you are in the intended account and workspace, then repeat the update steps. Buying credits does not add model access, and switching models does not restore a shared allowance. Ultrafast for Astra is limited at launch to Pro $500 and eligible Enterprise and Edu workspaces.
Go to the source
Sources & further reading
Based on the official documentation below. Features and access can change; check the provider before making a decision.
Was this page useful?
Help us make our coverage more useful. This is feedback on the page, not a rating of the tool.
Go a little further