OpenAI launches GPT-Rosalind-6 for biology research with gated access
Specialized model based on GPT-6 Astra enables dual-use biology queries for vetted researchers under trusted access controls.
What to know
- GPT-Rosalind-6 reuses GPT-6 Astra weights with modified access controls and safety framing for biology research, requiring applicant review before granting qualified access.
- Model shows significant performance gains over GPT-Rosalind-5.5 (49.6% vs 28.9% on MedChemBench) while maintaining refusals on biological weaponization requests.
- OpenAI classifies it as High in biological capability but Critical in cybersecurity, relying on vetting and governance practices rather than technical training changes.
The dispute Whether GPT-Rosalind-6 constitutes a distinct product or merely a repackaged base model with different access controls. · positions read across 9 posts and comments
The model appears to be rebranding existing weights with access controls rather than genuine technical innovation.
-
“So how is this different from Astra, if it's the same weights and zero other training has been done? Did they essentially package a system prompt?”
ludvigk · Hacker News ↗
OpenAI's terminology and product naming schemes lack clarity and consistency.
-
“The phrase system card is so frustrating because it's using the word card like it's a metaphor for something but then abuses it.”
deadnomore · Hacker News ↗
OpenAI AI developer
How it unfolded 1 development · click the chart to see its coverage articlespostscomments
-
1
OpenAI benchmarks GPT-Rosalind-6 against predecessor model
OpenAI evaluated GPT-Rosalind-6 on three life sciences research benchmarks and compared results with GPT-Rosalind-5.5. On MedChemBench v7, GPT-Rosalind-6 scored 49.6% versus 28.9% for the previous version. The model is rated High in biological and chemical capabilities, Critical in cybersecurity, and below High in AI self-improvement.
“GPT-Rosalind-6 is a frontier reasoning model for biology, drug discovery, and translational medicine.”
— OpenAI · source -
1 outlet GPT-Rosalind-6 System Card Addendum
first by HN Frontpage, 6d ago
-
Last I checked machine learning wasnt the bottleneck for drug discovery and where it fits in the research process was largely solved 5+ years ago.. Weren't there already a ton of ML based drug discovery startups that went out of business, or got aquired, not because they werent useful but because they produced more candidates than the industry…
2 more of the top 3 · 9 posts in this stretch
-
Slightly off-topic to this model card but has anyone noticed changes to GPT6 Astra about a week after 6.1 came out?I feel like 6 Astra just turned a lot lazier during planning and strangely enough claude has become a lot more assertive with Opus 5.5 while it used to be very lazy.It’s really obvious during planning, it used to be that claude would…
-
Great military terms: "dual use", "Preparedness Framework", etc.You just want to make all mathematicians and white collar employees unemployed and want to distract the press and the public with your extinction nonsense.Though you may care about bioweapons because they could affect you and your trafficked children.
-
-
background
OpenAI releases GPT-Rosalind-6 system card addendum — OpenAI published a system card addendum for GPT-Rosalind-6, a frontier reasoning model designed for biology, drug discovery, and translational medicine. The model uses the same weights as GPT-6 Astra with no additional training, and is made available only via trusted access to qualified researchers after review of their proposed uses and security controls.
What people are saying 5 voices from 1 site · best of 9 · verbatim
- Nov 2
-
So how is this different from Astra, if it's the same weights and zero other training has been done?Did they essentially package a system prompt?
-
The phrase system card is so frustrating because it’s using the word card like it’s a metaphor for something but then abuses it. A card is supposed to be small and concise. This has many, many pages and is an “addendum”.
-
OpenAI naming scheme makes no sense. I get naming is hard but atleast keep it consistent.
-
> GPT-Rosalind-6 uses the same weights as GPT-6 Astra, our flagship reasoning model.Didn’t Sol-6.1 top Astra? Or is it maybe that Astra is still a larger model and therefore can comprehend deeper complexity better in the long run?
-
It’s clear to me that open source will rule at the end of the day as everyone wants AI and nobody can rely on U.S crop/gov anymore.