OpenAI launched GPT-Rosalind-6 on October 30, a biology model available only to approved organizations and qualified researchers through its trusted-access Rosalind Research program.
The OpenAI system card says the model targets biology, drug discovery, and translational medicine. It uses the same weights as GPT-6 Astra, with no added training data or extra post-training for this release.
OpenAI reports 37.7% single-attempt success on GeneBench Pro v13 at the highest reasoning budget tested, versus 16.7% for GPT-Rosalind-5.5. The company also says that evaluation used a benchmark-specific configuration that differs from production.
For agent teams, access policy is the practical constraint: workflows need approval, qualified human review, and validation under the production configuration before outputs can guide experiments or consequential decisions.