GPT-Rosalind-6: OpenAI's gated biology reasoning model on GPT-6 Astra weights
OpenAI published GPT-Rosalind-6 on 30 October for biology, drug discovery and translational medicine, available only to vetted research organisations through a trusted-access programme. It reuses GPT-6 Astra's weights unchanged, with no new training data or post-training, and differs by drawing a looser refusal line on dual-use biology queries.
Why it matters
A frontier model repackaged as a controlled-access biology specialist: approved labs get fewer refusals on dual-use life-sciences work, at $15/$75 per million tokens, while the general model keeps refusing.
What you could build with it
A drug-discovery startup could route medicinal-chemistry and wet-lab troubleshooting queries through Rosalind-6, getting answers on dual-use-adjacent questions that the general model refuses, while keeping its compound data inside an approved workspace.
Does it hold up?
Too early to judge: access is gated to approved research customers only, with no public deployments reported yet, and the benchmark gains come from the Astra base rather than new biology training.
Built with GPT-Rosalind-6
- OpenAI's New Biology Model Has a Gate Before the Labarticle · Walks through the trusted-access gate, the Work/Codex-only availability, and what the model is actually for.
- OpenAI launches GPT-Rosalind-6 biology AI modelarticle · Reports the 30 October launch, the same-weights disclosure, benchmark gains over 5.5, and the High biological-risk rating.
- GPT-Rosalind-6: OpenAI's Gated Biology Model on GPT-6 Astra Weights, for Approved Customers Onlyarticle · Reads the system card addendum closely and stresses what the release is not: no new training, no public access, no clinical claim.
Learn more
First spotted on article: source.
More AI models
VoiceStudio: open-source, fully-local ElevenLabs alternative for voice cloning, dubbing, and audiobooksdebpalash/VoiceStudio is a desktop app that runs voice cloning, voice design, video dubbing, dictation, transcription…models · JEV 0.75Tencent Hunyuan Hy3: open 295B MoE with 21B active under Apache 2.0Tencent open-sourced Hy3, a 295B-parameter mixture-of-experts model with only 21B parameters active per token, under…models · JEV 0.73NVIDIA Isaac Lab 3.0 GA + Isaac OS open runtime + Agile ONE SN humanoidMajor Isaac platform expansion at GTC Berlin Oct 21: Isaac Lab 3.0 is generally available as the open-source robot…models · JEV 0.73NVIDIA Nemotron 3.5 Retriever collection: Embed 8B/1B, Rerank 8B, Parse 631MNVIDIA's Oct 21 collection drop for agentic retrieval: Nemotron-3.5-Embed in 8B BF16, 1B BF16 and 1B NVFP4…models · JEV 0.72Kimi K3.1 Pro: Moonshot's 2.8T open-weight flagship officially launchesMoonshot AI officially launched Kimi K3.1 Pro on Oct 23, a mixture-of-experts model with 2.8 trillion total / 104…models · JEV 0.72NVIDIA Cosmos 3.5 world models: now openly availableNVIDIA's Cosmos 3.5 family of open world foundation models for robots, autonomous vehicles and vision agents is now…models · JEV 0.71
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.