A leak spotted on X suggests OpenAI is testing a new subscription and experience called “ChatGPT for Science,” with references appearing in a web build. The catch: it may not be open to everyone, and could follow the same kind of eligibility and verification r
For the third time, it’s a science-focused ChatGPT—only this one comes wrapped in a subscription name.
A leak spotted on X points to a new offering called “ChatGPT for Science.” The feature’s references have also reportedly appeared in a web build. a detail that matters because it suggests OpenAI isn’t just toying with the idea in abstract terms. It’s being tried in something close to a real product experience.
Right now, OpenAI already offers multiple ways to use ChatGPT: ChatGPT for personal use, ChatGPT for Teams, and ChatGPT business/enterprise.
That existing lineup comes with clear access boundaries. Personal ChatGPT “works for everybody,” the source says. Teams requires users to have a company domain and at least three users. ChatGPT business is restricted to legal entities.
If ChatGPT for Science follows similar rules, the likely outcome is that only verified institutes or universities would be able to use it—rather than researchers signing up as individuals.
That idea sits inside a broader push from OpenAI to tailor models and subscriptions for scientific workflows. It’s not the first time the company has tried to build an offering specifically for science use cases.
Earlier, OpenAI announced GPT-Rosalind, described as being built on the foundation of its advanced GPT-5.5 architecture. The source says GPT-Rosalind isn’t simply a reskinned ChatGPT with a science prompt. Instead, it’s positioned as a highly specialized, purpose-built model for enterprise-scale life sciences research.
Access is where the story turns. GPT-Rosalind is locked behind what OpenAI calls a “trusted-access deployment structure.” In practice. that means it’s strictly available to eligible organizations—such as major pharmaceutical companies like Novo Nordisk or verified research institutions—conducting legitimate. public-benefit scientific research.
The package also comes with enterprise-grade security and strong safety governance, described as mirroring and exceeding the strict requirements of ChatGPT Enterprise.
The leak’s bigger implication is how OpenAI might take some of that capability and widen the funnel. The source suggests OpenAI could be planning to bring at least some of these capabilities to all institutions through ChatGPT for Science, instead of keeping them locked to select partners.
If that’s the direction, ChatGPT for Science would likely offer more than a standard subscription with a science-friendly interface. The source describes it as having special grounding in discoveries and research around scientific topics compared to a regular subscription.
At the moment, there’s no confirmed launch date. What’s clear is that the product is actively being tested on the web build. and the source adds that an announcement is likely “weeks away.” For now. the unanswered question is just as sharp as the promise: will this science-focused experience reach researchers broadly. or will it stay behind verification gates like the rest of OpenAI’s higher-control offerings?.
OpenAI ChatGPT for Science ChatGPT subscription GPT-Rosalind AI for science life sciences research OpenAI Teams ChatGPT Enterprise trusted-access deployment structure