Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Open ChatGPT’s model picker, then choose any model or reasoning level displayed for your account. The list is account-specific: your plan, rollout status, workspace permissions, model limits and whether a model is still supported all affect what appears. If an option is missing, the picker and your workspace settings are more reliable than an old screenshot or a universal model list.
Switch models in ChatGPT
ChatGPT changes models from the model picker attached to a conversation. The exact placement and labels can change as the interface is updated, so look for the control that names the current model or selection rather than following a fixed screenshot.
- Open ChatGPT and enter the conversation where you want to change the model.
- Open the model picker shown for that account. It may be positioned near the conversation header or composer, depending on the current interface.
- Review the models and reasoning choices displayed in the picker. Only choices currently eligible for that account can be selected.
- Select the model or reasoning level you want.
- Send a new message and check the conversation’s active label if you need to confirm the selection.
A model change applies to the conversation or message flow where you make it; it does not grant access to models that the account, plan or workspace has not enabled. If the picker offers a generic Latest choice and your account is eligible for the current GPT-6 experience, the current guide says GPT-6 appears in place of that generic selection. An explicit GPT-5.6 Sol selection remains on GPT-5.6 rather than silently becoming the generic latest option.
Choose speed or deeper reasoning
Some accounts expose reasoning controls separately from the model name. The current guidance describes these levels:
#1 Best Overall
| Choice | What it is for | Availability |
|---|---|---|
| Instant | Fast responses when latency matters more than extended reasoning. | Plan-dependent; it appears only when enabled for the account. |
| Medium | A balanced setting for ordinary questions that need more thought than a rapid response. | Plan-dependent. |
| High | More reasoning depth for difficult analysis, multi-step work or ambiguous requests. | Plan-dependent. |
| Extra High | The deepest listed reasoning setting for accounts that expose it. | Restricted to eligible plans or workspaces. |
| Pro | A high-end reasoning choice shown only to accounts that qualify for it. | Plan-dependent. |
| Think | A harder-question mode available to Free and Go users under the current guidance. | Available to Free and Go users when shown in their picker. |
Use the fastest option that reliably handles the task. Instant is sensible for short lookups, rewriting and routine questions. Medium or High is more appropriate when the answer requires several constraints, calculations or a careful comparison. Extra High or Pro can be useful for especially demanding work, but those settings can consume an allowance differently from faster choices. The picker is the authority for the choices and labels on your account.
Why a model is missing
A missing entry does not by itself indicate a malfunction. Four conditions commonly determine whether a choice appears:
- Plan eligibility: Free, Go and paid plans do not receive identical model or reasoning choices. In particular, Free and Go users do not receive the same GPT-5.6 Sol options as eligible paid plans.
- Rollout status: A feature can be released progressively. Two accounts on the same nominal plan may temporarily see different choices while a rollout is in progress.
- Workspace permissions: In a managed organization, an administrator can restrict which models members may use. Your personal preference cannot override that policy.
- Support status: A retired model is removed from the picker. It cannot be restored by changing a local setting.
Because these conditions change, do not treat a third-party list or an old screen capture as a promise that your account should contain a particular model. OpenAI’s current business guidance says users should check the model picker for the options on their account because availability and limits change over time.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Check your plan and workspace before troubleshooting
Personal accounts
Open the account or subscription area in ChatGPT and confirm the plan associated with the signed-in account. A user who is signed into a different account, or who expects a paid-plan option while using a Free or Go account, can see a picker that is working as designed. The plan name alone still does not guarantee immediate access when a feature is rolling out.
Managed workspaces
If the conversation is inside a company, school or other managed workspace, ask an administrator which models are enabled for members. Workspace controls can remove a model from the picker even when that model exists elsewhere. An administrator’s setting also explains why colleagues may see a different list from yours.
Separate ChatGPT from the API
ChatGPT model-picker access and API model availability are separate matters. A model visible in ChatGPT is not automatically available to an API project, and an API entitlement does not add a model to the ChatGPT picker. Check the product you actually intend to use before drawing conclusions from the other one.
What happens when you reach a usage limit
Allowances depend on both your plan and the model or reasoning choice. When an allowance is exhausted, ChatGPT can continue with another available model instead of stopping the conversation. If a reset time is available, the interface displays it; wait until that time if you need the original choice restored.
- Fallback: The conversation may continue on another model that remains available to your account.
- Reset: ChatGPT can show when the affected allowance becomes available again.
- Shared allowance: Switching models is not a guaranteed way to obtain more usage. Models may draw from the same limit, so changing the picker selection may not increase the remaining allowance.
- Different limits: A faster model, a deeper reasoning level and a legacy model can have different eligibility or allowance rules; rely on the notice shown in the product rather than a remembered cap.
If quality changes after a limit message, inspect the active model label before assuming the original model is still answering. Record the displayed reset time if the task is important, and avoid repeatedly switching choices as a strategy for bypassing a shared allowance.
Use legacy models when your workspace allows them
Legacy access is conditional rather than permanent. For Enterprise and Edu workspaces, an administrator can enable legacy model access in workspace settings. A model then appears only while both conditions remain true:
- Legacy model access is enabled by the administrator.
- The particular model is still available from the service.
OpenAI states: “A legacy model appears in the picker only while it remains available and legacy model access is enabled.” If either condition fails, the model is absent. Retired models no longer appear in the picker, and existing conversations may be moved to a newer supported model. Enabling the setting cannot revive a model that has been retired.
For an Enterprise or Edu user who cannot see a needed older model, the practical sequence is to ask the administrator to verify the legacy-access setting, confirm that the model has not been retired, and then reopen the picker. This does not change API availability, which is managed separately.
Free tools Windows power users keep installed
One-click scans. No signup required.
A practical method for selecting a model
For a quick, low-risk task
Choose Instant when the job is a short answer, basic transformation or routine drafting request. Confirm the result directly instead of selecting a deeper mode solely because it is available.
For a multi-step question
Choose Medium or High when the task contains several requirements, competing constraints, calculations or a decision that benefits from explicit reasoning. If the picker exposes Extra High or Pro and the task justifies the additional allowance, those options provide greater reasoning depth.
For Free or Go accounts
Use Think when it is displayed and the question is unusually difficult. Do not expect the same GPT-5.6 Sol choices offered to eligible paid plans; the absence is a plan distinction, not necessarily an account error.
For a compatibility-sensitive workflow
Use the explicit model name shown in the picker instead of a generic Latest choice when you need to keep a workflow tied to a particular current model. Remember that an explicit name remains subject to retirement, workspace policy and usage limits.
Troubleshoot a missing model
Work through these checks in order so you can distinguish eligibility from a temporary interface problem.
- Confirm the signed-in account. Check that you are using the account and workspace associated with the expected plan.
- Reopen the picker. Close and reopen the model menu, then start a new conversation if the current thread has retained an older selection.
- Read the plan entitlement. Compare the displayed plan with the model’s eligibility. Free and Go do not expose every GPT-5.6 Sol option available to eligible paid plans.
- Check for a rollout difference. If a colleague sees the option but you do not, a staged rollout may be responsible even when both accounts appear similar.
- Ask the workspace administrator. For managed accounts, verify that policy has not limited the available models or reasoning levels.
- Check whether the model is retired. A retired model will not return to the picker through a user-side setting.
- Inspect limit notices. If the model appeared earlier but is now unavailable, read the usage message and displayed reset time. The service may have moved the conversation to a fallback model.
- Compare product surfaces. Do not use API documentation or API access as proof that the ChatGPT picker should contain the same model.
If these checks do not explain the difference, save the exact model label, account plan, workspace type and message shown in the interface before contacting support or an administrator. Those details are more useful than describing an option from memory.
How to manage availability in a team
Administrators should treat model availability as a policy and change-management issue rather than a static menu. Before enabling a legacy model or restricting current choices, document which teams need it, what happens when its allowance is exhausted, and what replacement users should select after retirement.
- Tell members which workspace they must open for the approved picker to appear.
- Explain whether users should choose a named model, a reasoning level or the generic Latest option for ordinary work.
- Warn that a model can disappear when it is retired even if a conversation previously used it.
- Provide a fallback instruction for limit events, including where members can see a reset time.
- Review policy after product updates because rollout status, limits and model names can change.
Members should not attempt to evade a workspace restriction by switching accounts or repeatedly changing models. The correct resolution is an administrator decision or a different eligible workspace.
Recommended Free Tools
Common mistakes to avoid
Assuming every account has the same picker
Plan, rollout and workspace differences make a universal list misleading. Use the controls visible in the account you are operating.
Best Value
Confusing a fallback with a model switch you requested
After a limit, ChatGPT may continue with another available model. Check the active label before comparing answers or diagnosing a quality change.
Expecting a switch to reset usage
If models share an allowance, selecting another one does not create additional capacity. Wait for the displayed reset or use the fallback that remains available.
Expecting legacy access to be permanent
An administrator must enable legacy access, and the model must still be offered. Either condition can change.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchUsing old screenshots as current documentation
Labels, model names and reasoning choices are subject to change. A current picker is a more accurate account-level reference than a static image.
Decision table: which choice should you make?
| Your situation | First choice | Check before sending |
|---|---|---|
| Short response or routine rewrite | Instant, if shown | Confirm the task does not require extended reasoning. |
| Several constraints or a detailed comparison | Medium or High | Check the remaining allowance and expected response time. |
| Very difficult analysis and an eligible paid plan | Extra High or Pro, if shown | Confirm that the additional reasoning choice is permitted in the current workspace. |
| Free or Go account facing a hard question | Think, if shown | Do not expect the GPT-5.6 Sol options available to eligible paid plans. |
| Need to stay on a named current model | Select its explicit name | Remember that limits, policy and retirement still apply. |
| Need an older model in Enterprise or Edu | Use the legacy entry, if shown | Ask an administrator to enable legacy access and verify the model remains available. |
Keep your account-specific instructions current
Availability is a moving product setting, not a permanent specification. When a model matters to your work, record the exact label shown in the picker, the workspace in which it appears and any limit or reset notice. Recheck those details after a plan change, workspace-policy change or major ChatGPT update.
The safest operating rule is simple: choose from the picker you actually have, treat the displayed allowance and reset notice as authoritative, and ask the workspace administrator about anything controlled centrally. This approach remains valid even when model names, reasoning levels or rollout groups change.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
