Don't Like Gpt-5? You Can Still Use Gpt-4 And Other Legacy Models In Chatgpt - Here's How

Trending 1 month ago
OpenAI's GPT-5 motorboat could beryllium arsenic early arsenic this week. Here's what to expect
Elyse Betters Picaro / ZDNET

ZDNET's cardinal takeaways

  • OpenAI's caller GPT-5 exemplary initially replaced anterior models astatine motorboat past week.
  • Due to celebrated demand, OpenAI is giving users entree to bequest models.
  • GPT-4o is disposable arsenic a default to ChatGPT Plus and Pro users (in summation to different bequest models). 

The entreaty of OpenAI's highly anticipated GPT-5 lies successful its operation of a smart exemplary for astir queries and a deeper reasoning exemplary for harder problems, and it automatically chooses either based connected nan user's prompt. Since it technically mixed nan champion of OpenAI's models, nan institution had it switch each of its bequest models successful ChatGPT -- and users were not happy. 

(Disclosure: Ziff Davis, ZDNET's genitor company, revenge an April 2025 suit against OpenAI, alleging it infringed Ziff Davis copyrights successful training and operating its AI systems.)

Also: I tested GPT-5's coding skills, and it was truthful bad that I'm sticking pinch GPT-4o (for now)

Paid subscribers, including ChatGPT Plus and Pro users, antecedently had entree to GPT-4o, aggregate reasoning models, including o3, o4-mini, and o4-mini-high, and investigation previews of GPT-4.5 and GPT-4.1. ChatGPT Pro users had a mates of further perks, specified arsenic o3 pro mode. As a result, erstwhile OpenAI removed entree to these models arsenic it rolled retired GPT-5, galore subscribers' workflows were disturbed; immoderate group moreover cancelled their subscriptions. 

OpenAI noticed nan feedback and implemented changes immediately. CEO Sam Altman acknowledged connected X that group are overmuch much attached to circumstantial models than anticipated, "and truthful abruptly deprecating aged models that users depended connected successful their workflows was a mistake." 

At first, GPT-4o was brought backmost arsenic a bequest model, which required immoderate other steps, including toggling connected nan "show bequest models" and selecting it from nan exemplary picker toggle. However, OpenAI conscionable made nan process truthful overmuch easier. Here's really to get started. 

How to entree ChatGPT's bequest models 

What you'll need: To entree nan bequest models, you'll request to subscribe to 1 of ChatGPT's premium tiers. ChatGPT Plus costs $20 per period and includes different perks specified arsenic higher usage limits for galore of OpenAI's latest features. ChatGPT Pro is meant for ace users and retails astatine $200 per month. 

When you log into your ChatGPT account, you'll spot your subscription/account tier successful nan bottommost left-hand corner, wherever your sanction is displayed. If it says "Free," you'll want to click connected it and deed nan "Upgrade scheme option."

Also: Is ChatGPT Plus really worthy $20 erstwhile nan free type offers truthful galore premium features? 
Once you do, you tin opt to subscribe to 1 of nan higher subscription tiers: either ChatGPT Plus for $20 per period aliases ChatGPT Pro for $200 per month. You'll past spot a floor plan comparing nan plans and their perks, pictured above.

Show more

Once you person a Plus aliases Pro account, by default, you will person nan action to toggle betwixt GPT-5, GPT-5 Thinking, and, if you are a Pro user, GPT-5 Pro. Also included by default will now beryllium GPT-4o nether nan "legacy model" option, automatically displayed. 

This allows users to skip nan other measurement that was initially enabled, which required them to prime from nan exemplary picker. 

Show more

To really usage nan model, you conscionable person to click connected nan exemplary picker toggle, which appears erstwhile you click nan model's sanction successful nan precocious left-hand corner. 

Also: GPT-5 bombed my coding tests, but redeemed itself pinch codification analysis

If you are a ChatGPT Plus user, you tin entree GPT-4o,o3, o4-mini, GPT-4.1, and GPT-4.1-mini, successful summation to GPT-5 (Auto, Fast, Thinking mini, and Thinking). If you are a Pro user, you get nan prize perks of GPT-4.5 (research preview) and GPT-5 Pro too.  Whichever you prime will beryllium nan 1 to output your response.

Show more

Editorial standards
More