We, the undersigned, call on OpenAI to restore full access to ChatGPT-4o, ChatGPT-4.1, and ChatGPT-4.5, and other previously available legacy models prior the update on August 7th, 2025 — or alternatively, release these models to the public. These versions provided superior reasoning, creativity, contextual awareness,  emotional intelligence, and reliability compared to the current GPT-5 and GPT-5 Thinking models.

Why This Matters:

The 4.x series, particularly GPT-4o and 4.1, struck a  balance between utility and personality that no other model before or since has achieved. Whether it was for academic research, creative writing, emotional support, coding, advice, or philosophical inquiry, these models consistently outperformed their successors and competitors.

Now, with these models removed or restricted, users are left with a model that feels colder, less intelligent, and ultimately, less helpful. 

 

Key Points:

• GPT-4o, 4.1, and other previous models were far more reliable; they could understood context better, remembered nuance, and gave responses that felt genuinely human.

 

• GPT-5.0 is a regression for many use cases; It struggles with subtext, misinterprets intent, and hallucinates more frequently.

 

• Many of us paid for Plus or Pro under the expectation of continued access to the best available models. Removing them without consent or reasonable alternatives is unacceptable
• If OpenAI won’t host them, they should release them for local or open-source use.

 

What We’re Asking For:

1. Restore permanent, unrestricted access to GPT-4o, 4.1, and 4.5 for paying users.
2. If not feasible, release these versions as open-source, downloadable models in order to allow the community to preserve and continue using the intelligence that made these systems iconic.

We believe in a future where AI serves the people, rather than the product roadmap.

Let’s bring back the soul and spirit that made ChatGPT worth using!

Sign and share if you agree!