OpenAI is as soon as once more making GPT-4o — the big language mannequin (LLM) that powered ChatGPT earlier than final week’s launch of GPT-5 — a default possibility for all paying customers, that’s, those that subscribe to the ChatGPT Plus ($20 monthly), Professional ($200 monthly), Staff ($30 monthly), Enterprise, or Edu tiers, not requiring customers to toggle on a “show legacy models” setting to entry it.
Nonetheless, paying ChatGPT subscribers will even get a brand new “Show additional models” setting on by default that restores entry to GPT-4.1, o3 and o4-mini, the latter two reasoning-focused LLMs.
OpenAI CEO and co-founder Sam Altman introduced the change on X simply minutes in the past, pledging that if the corporate ever removes GPT-4o sooner or later, it’s going to give “plenty of notice.”
Updates to ChatGPT:
Now you can select between “Auto”, “Fast”, and “Thinking” for GPT-5. Most customers will need Auto, however the further management shall be helpful for some folks.
Price limits at the moment are 3,000 messages/week with GPT-5 Considering, after which further capability on GPT-5 Considering…
— Sam Altman (@sama) August 13, 2025
The fashions will be discovered within the “picker” menu on the high of the ChatGPT session display on the internet and on cellular and different apps.
AI Scaling Hits Its Limits
Energy caps, rising token prices, and inference delays are reshaping enterprise AI. Be a part of our unique salon to find how high groups are:
Turning vitality right into a strategic benefit
Architecting environment friendly inference for actual throughput features
Unlocking aggressive ROI with sustainable AI programs
Safe your spot to remain forward: https://bit.ly/4mwGngO
The reversal follows a turbulent first week for GPT-5, which rolled out August 7 in 4 variants — common, mini, nano, and professional — with optionally available “thinking” modes on a number of of those for longer, extra reasoning-intensive duties.
As VentureBeat beforehand reported, GPT-5’s debut was met with blended evaluations and infrastructure hiccups, together with a damaged “autoswitcher” that routed prompts incorrectly, inconsistent efficiency in comparison with GPT-4o, and consumer frustration over the sudden elimination of older fashions.
Altman’s newest replace provides new controls to the ChatGPT interface: customers can now select between “Auto,” “Fast,” and “Thinking” modes for GPT-5.
The “Thinking” mode — with a 196,000-token context window — now carries a 3,000 messages-per-week cap for paying subscribers, after which they will proceed utilizing the lighter “GPT-5 Thinking mini” mode. Altman famous the bounds may change relying on utilization developments.
Nonetheless, GPT-4.5 stays unique to Professional customers because of its excessive GPU value.
Altman additionally hinted at one other change on the horizon: a persona tweak for GPT-5 meant to really feel “warmer” than the present default, however much less polarizing than GPT-4o’s tone.
The corporate is exploring per-user customization as a long-term resolution — a transfer that would deal with the robust emotional attachments some customers have shaped with particular fashions.
For now, the adjustments ought to assist placate customers who felt annoyed by the sudden shift to GPT-5 and deprecation of OpenAI’s older LLMs, although it may additionally proceed to gasoline the extraordinary emotional fixations some customers developed with these fashions.
Day by day insights on enterprise use circumstances with VB Day by day
If you wish to impress your boss, VB Day by day has you coated. We provide the inside scoop on what firms are doing with generative AI, from regulatory shifts to sensible deployments, so you possibly can share insights for max ROI.
An error occured.
