A new leak spotted on X indicates that OpenAI is testing a potential new subscription tier called “ChatGPT for Science,” designed specifically for scientific and research-focused use cases.
The references were reportedly found in the web build of ChatGPT, suggesting that the feature is still under development and not yet publicly available.
At present, OpenAI offers ChatGPT in multiple tiers, including personal use, Teams, and enterprise-level subscriptions.
While the personal version is available to all users, Teams requires a company domain and a minimum number of users, and enterprise access is typically restricted to verified legal entities.
Early indications suggest that “ChatGPT for Science” could follow a similar model, potentially limiting access to verified universities, laboratories, and research institutions.
OpenAI’s growing focus on Scientific AI
This is not the first time OpenAI has explored tools designed specifically for scientific applications.
The company recently introduced GPT-Rosalind, a specialized model built on its advanced GPT-5.5 architecture. Unlike general-purpose ChatGPT, GPT-Rosalind is designed for life sciences and large-scale research environments.
According to available details, GPT-Rosalind operates under a “trusted-access deployment structure,” meaning it is only available to select organizations such as pharmaceutical companies and approved research institutions.
Potential expansion of research tools
Reports suggest that OpenAI may use “ChatGPT for Science” to expand access to advanced research capabilities beyond a small group of partners.
Instead of limiting tools like GPT-Rosalind to select institutions, the new subscription could make similar scientific features more widely available to eligible academic and research organizations.
The goal appears to be strengthening ChatGPT’s grounding in scientific literature, discovery, and research workflows compared to standard consumer versions.
While the feature has been spotted in testing, OpenAI has not officially announced “ChatGPT for Science,” and no release date has been confirmed.
However, the presence of early references suggests that an announcement could arrive in the coming weeks if development continues at the current pace.
