> ## Content Index
> Fetch the complete content index at: https://www.testingcatalog.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# OpenAI lists new gpt-rosalind-discovery model in docs
- URL: https://www.testingcatalog.com/openai-lists-new-gpt-rosalind-discovery-model-in-docs/
- Published: 2026-10-10T14:57:48.000Z
- Updated: 2026-10-10T14:57:48.000Z
- Description: OpenAI's API pricing page now lists gpt-rosalind-discovery, a life sciences model the company hasn't announced. It costs the same as gpt-rosalind-research.
- Author: Dwaybe Cubbins
- Tags: OpenAI, Latest AI News, AI Rumours

OpenAI's [API pricing page](https://developers.openai.com/api/docs/pricing?ref=testingcatalog.com) seems to have been updated to include a new model the company hasn't officially announced yet. Under a section called "Life sciences models", **gpt-rosalind-discovery** now sits right beside gpt-rosalind-research. The listing was first shared by the [AI Tracker Bot](https://x.com/aitrackerbot/status/2108795714880876777?ref=testingcatalog.com) on X. There's still no blog post, help article, or system card for it yet.

The pricing is interesting because both models cost the same: $5 per million input tokens, $0.50 per million cached input tokens, and $25 per million output tokens. The page states that billing for **gpt-rosalind-research** began on October 5, 2026, and that cache-write pricing doesn't apply. Access remains restricted to approved internal research via OpenAI's trusted-access program. Data residency endpoints add a 10% uplift, covering eligible models released on or after March 5, 2026.

Another factor to note is that OpenAI says *"all eligible organizations will continue to get access to the latest GPT-Rosalind models as they're released."* That makes it clear that the company had aimed for a family of models, not a single release, and a *"discovery"* variant fits that plan.

> OPENAI 🔥: A new gpt-rosalind-discovery model has been spotted on the API pricing page. The model hasn’t been publicly announced so far.   
>  
> The model has the same pricing as gpt-rosalind-research while its exact purpose is yet unclear.   
>  
> Should it be specifically a drug discovery… [https://t.co/wjVAT2aJwq](https://t.co/wjVAT2aJwq?ref=testingcatalog.com) [pic.twitter.com/m4cM9VOvlS](https://t.co/m4cM9VOvlS?ref=testingcatalog.com)
> 
> — 🚨 AI News | TestingCatalog (@testingcatalog) [October 10, 2026](https://x.com/testingcatalog/status/2108804914071658577?ref%5Fsrc=twsrc%5Etfw&ref=testingcatalog.com)

While the company hasn't exactly said what **gpt-rosalind-discovery** is, the name suggests it could be a drug discovery model. It's not surprising because when OpenAI [introduced GPT-Rosalind](https://openai.com/index/introducing-gpt-rosalind/?ref=testingcatalog.com) in April, the company called it a frontier reasoning model for biology, drug discovery and translational medicine, and named Amgen, Moderna, the Allen Institute and Thermo Fisher Scientific among its early partners. The help center says the current model is most useful for early discovery and hypothesis generation, such as target biology, literature synthesis, and omics interpretation. A model built specifically for target identification or molecular design would fit OpenAI's plans. For now, though, that's a guess.

For context, OpenAI removed GPT-Rosalind from research preview on September 11 and opened it to eligible organizations worldwide through the trusted-access program. Published pricing took effect on October 5\. The latest version, GPT-Rosalind-5.5, is trained on top of GPT-5.5\. According to its [system card](https://deploymentsafety.openai.com/gpt-rosalind-5-5/trust-based-access?ref=testingcatalog.com), it's trained not to refuse sophisticated biology questions, and gated access is the main safeguard instead. OpenAI took a similar approach to security when it [expanded trusted access for GPT-5.4-Cyber](https://www.testingcatalog.com/openai-expands-trusted-access-for-gpt-5-4-cyber/). That seems to be how OpenAI releases specialized models that could be misused.

That said, OpenAI isn't the only company advancing in life sciences. Anthropic has been working on [Claude Operon, a biology-focused mode in the Claude desktop app](https://www.testingcatalog.com/anthropic-tests-claude-operon-for-scientific-research-in-biology/), and we recently got an [early look at Claude Science](https://www.testingcatalog.com/early-look-at-anthropics-claude-science-app-for-researchers/), which is [currently in beta](https://claude.com/product/claude-science?ref=testingcatalog.com). OpenAI also has its own ChatGPT for Academic Researchers project. Together with Rosalind, that would give OpenAI a fairly complete offering for research labs.

OpenAI often lists models on its pricing page shortly before announcing them, so gpt-rosalind-discovery may not stay quiet for long. When it's announced, the main questions will be whether it's a separate model trained for discovery work or a differently tuned version of the research model, and whether the trusted-access requirements stay the same.