OpenAI appears to be testing a new subscription and experience for science use cases, but it’s unclear if it’ll be available to everyone regardless of their background.
As spotted on X, this new subscription/model is called “ChatGPT for Science,” and references to the feature were spotted on the web build.
Right now, OpenAI offers ChatGPT for personal use, Teams, and business/enterprise.
While ChatGPT personal works for everybody, Teams requires you to have a company domain and at least three users. On the other hand, ChatGPT business is restricted to legal entities.
It’s likely that ChatGPT for Science will have similar restrictions, and only verified institutes or universities would be able to use it.
OpenAI has been optimizing ChatGPT for scientific use cases
It’s not the first time we’ve seen OpenAI’s attempt to build models or subscriptions specifically for science use cases.
OpenAI recently announced GPT-Rosalind, which is built on the foundation of its advanced GPT-5.5 architecture, but it’s not just a reskinned ChatGPT with a science prompt.
Instead, it is a highly specialized, purpose-built model designed specifically for enterprise-scale life sciences research.
GPT-Rosalind is locked behind what OpenAI calls a “trusted-access deployment structure.”
This means the model is strictly available to eligible organizations, such as major pharmaceutical companies like Novo Nordisk or verified research institutions, that are conducting legitimate, public-benefit scientific research.
It requires enterprise-grade security and strong safety governance, mirroring and exceeding the strict requirements of ChatGPT Enterprise.
OpenAI may be planning to bring some of these capabilities to all institutions through ChatGPT for Science, rather than locking them to select partners.
In other words, ChatGPT for Science will have special grounding in discoveries and research around scientific topics compared to a regular subscription.
At the moment, we don’t know when ChatGPT for Science will go live, but it’s being actively tested on the web, and an announcement is likely weeks away.
Security teams log 54% of successful attacks and alert on just 14%. The rest move through your environment unseen.
The Picus whitepaper shows how breach and attack simulation tests your SIEM and EDR rules so threats stop slipping by detection.
