The frontier reasoning model shares its foundational weights with GPT-6 Astra but is specifically tuned to serve qualified researchers in biology and drug development. While it permits deeper exploration of dual-use biological concepts, the system card addendum emphasizes that requests involving weaponization remain blocked under rigorous deployment safeguards. Evaluated under the developer's Preparedness Framework, the model reached a High capability rating in biological and chemical risks, avoiding a Critical designation. For non-biological domains, the system defers to the parent model's ratings, which include a Critical assessment in Cybersecurity.
Key points
Researchers receive specialized biological reasoning tools while safeguards continue to p…: Researchers receive specialized biological reasoning tools while safeguards continue to prohibit content related to biological weaponization.
Applicants must satisfy strict governance and security requirements under a trusted acces…: Applicants must satisfy strict governance and security requirements under a trusted access framework before utilizing the model.
The parent model GPT-6 Astra retains a Critical capability rating in Cybersecurity, which…: The parent model GPT-6 Astra retains a Critical capability rating in Cybersecurity, which applies directly to this release.
Sources checked by our analysis pipeline: 1
GPT-Rosalind-6 is a frontier reasoning model for biology, drug discovery, and translational medicine. We offer it as GPT-Rosalind-6 Research to qualified scientists, research institutes, companies and government partners. Rosalind has fewer refusals on dual-use biological queries but is only made available via trusted access. We review applicants’ proposed uses, security controls and governance practices before granting access. We also continue to have other deployment safeguards in place.
GPT-Rosalind-6 uses the same weights as GPT-6 Astra, our flagship reasoning model. It is designed to support legitimate biological research, including some dual-use topics, while still refusing requests for biological weaponization and retaining refusals on other safety topics.
Under our Preparedness Framework, we are treating this launch as High capability in the Biological and Chemical domain, below the threshold for Critical. We rely on GPT-6 Astra’s capability assessments for other Preparedness risks and treat GPT-Rosalind-6 as Critical in Cybersecurity and below High in AI Self-Improvement.
This addendum describes GPT-Rosalind-6’s biological capability assessment, additional safety and alignment evaluations, and deployment safeguards. We include results for GPT-Rosalind-5.5 where available for comparison. This addendum should be…
Users of the platform can now assemble bespoke tools directly within the system to fit specialized multimedia needs. This functionality addresses the demand for tailored workflows by supporting advanced applications like rendering and video manipulation. The update significantly broadens what operators can produce without leaving the native environment.
The artificial intelligence company has officially submitted its initial public offering documentation to launch a public listing. This move marks a critical step forward in the company's corporate development and transition to public market scrutiny. Market participants are watching the submission closely as a benchmark for broader industry momentum and investor sentiment.
Frontier generative models have drastically lowered the barrier to entry for game development by allowing creators to generate instant browser experiences with code-driven audio. However, turning these simple prototypes into viable commercial products remains difficult due to architectural limitations in complex logic and persistent multiplayer systems. At the same time, rising player backlash and platform disclosure mandates are forcing studios to keep human specialists central to the process.