OpenAI tests ChatGPT for Science subscription with restricted access

At a glance:

• OpenAI is testing a new subscription tier called "ChatGPT for Science" for scientific research use cases
• Access likely restricted to verified institutions and universities, mirroring enterprise-level restrictions
• Builds on GPT-Rosalind model designed for life sciences research with strict safety protocols

What's the new subscription

OpenAI appears to be developing a specialized version of ChatGPT tailored for scientific research, dubbed "ChatGPT for Science." References to this subscription were spotted in the web build of ChatGPT, suggesting active development. The company currently offers three main ChatGPT tiers: personal use, Teams for businesses with company domains and three or more users, and business/enterprise plans restricted to legal entities. The science-focused subscription is expected to follow a similar restricted access model, potentially limiting availability to accredited academic or research institutions rather than general public access.

How it compares to existing offerings

Unlike the open-access personal ChatGPT, the science variant may require institutional verification to ensure responsible use of advanced capabilities. This approach aligns with OpenAI's recent launch of GPT-Rosalind, a model built on GPT-5.5 architecture specifically for enterprise-scale life sciences research. GPT-Rosalind operates under a "trusted-access deployment structure," available only to organizations like Novo Nordisk or verified research institutions conducting public-benefit science. The new subscription could democratize some of these capabilities while maintaining safety standards.

GPT-Rosalind's role in the strategy

GPT-Rosalind represents OpenAI's deeper push into specialized AI applications. Unlike generic ChatGPT with science prompts, it is a purpose-built model optimized for complex life sciences tasks. Its deployment requires enterprise-grade security and governance frameworks that exceed standard ChatGPT Enterprise safeguards. By creating a dedicated science subscription, OpenAI may aim to bridge the gap between exclusive partnerships and broader institutional access, enabling more researchers to leverage advanced tools while preserving oversight.

What to watch next

The timeline for ChatGPT for Science remains unclear, though active web testing suggests an imminent announcement. Key questions include pricing models, specific features tailored for scientific workflows, and whether access will eventually expand beyond verified institutions. The move reflects growing demand for AI in research, with potential implications for pharmaceutical development, academic studies, and collaborative scientific projects. Investors and researchers alike will be watching for details on how this subscription integrates with OpenAI's broader enterprise strategy and safety commitments.